article
stringlengths 0
745k
| abstract
stringlengths 0
92.7k
|
|---|---|
study of ce - intermetallics have drawn significant attention during past few decades due to the observation of many unusual properties such as valence fluctuations , kondo screening , heavy fermion superconductivity in these systems .
such properties arise due to the proximity of ce 4@xmath3 level to the fermi level leading to strong hybridization between the ce 4@xmath3 states and the conduction electronic states @xcite .
such hybridization often leads to a logarithmic enhancement of electrical resistivity at low temperatures in contrast to a decrease expected in a metal .
this is known as kondo effect . in the case of strong kondo coupling , the antiparallel coupling of the ce moment with the conduction electrons forms a singlet ground state called kondo singlet that manifests as a sharp feature ( kondo resonance feature ) in the electronic structure in the vicinity of the fermi level , @xmath6 .
a lot of success has been achieved to describe these systems within the anderson impurity models . here , the parameters defining the hybridization between 4@xmath3 states and valence electronic states are often estimated using band structure calculations based on local density approximations ( lda)@xcite .
the scenario can be different if these materials contain transition metals : the @xmath2 electrons forming the conduction band are strongly correlated .
the finite correlation strength among them suggests non - applicability of the band structure results within lda to derive the hybridization parameters and indicate the need to go beyond lda prescriptions .
a kondo lattice compound , ce@xmath0cosi@xmath1 is a good candidate for this study .
ce@xmath0cosi@xmath1 crystallizes in a alb@xmath0 derived hexagonal structure ( space group @xmath7 ) and is a mixed valent ( kondo lattice ) compound @xcite .
the electrical transport measurements revealed temperature dependence@xcite typical of a mixed valent system @xcite .
no signature of magnetic ordering was observed in the magnetic susceptibility measurements down to 0.5 k @xcite .
interestingly , gradual substitution of rh at co sites leads to plethora of interesting features due to increasing dominance of indirect exchange interaction @xcite .
for example , @xmath8 = 0.6 composition in ce@xmath0rh@xmath9co@xmath10si@xmath1 exhibits quantum critical behavior .
intermediate compositions having higher rh concentration exhibit signature of spin density wave ( sdw ) state @xcite .
clearly , the @xmath2 electronic states corresponding to the transition metals plays a key role in determining the electronic properties in this interesting class of compounds .
it is thus , important to probe the role of electron correlation in the electronic structure of these compounds . in this paper
, we report our results of investigation of the electronic structure of ce@xmath0cosi@xmath1 using high resolution photoemission spectroscopy and _ ab initio _ band structure calculations .
high energy resolution employed in our measurements enabled us to reveal kondo - resonance feature and the corresponding spin orbit satellite .
the comparison of the experimental spectra and the calculated ones indicate that the correlation strength among co 3@xmath2 electrons is significant ( @xmath11 3 ev ) .
the contribution of the co 3@xmath2 partial density of states ( pdos ) is small in the vicinity of the fermi level , where ce 4@xmath3 contributions are dominant .
ce@xmath0cosi@xmath1 was prepared by melting together stoichiometric amounts of high purity ( @xmath12 99.9% ) ce , co and si in an arc furnace .
the single phase was confirmed by the absence of impurity peaks in the @xmath8-ray diffraction pattern .
the specimen was further characterized by scanning electron microscopic measurements and energy dispersive @xmath8-ray analysis @xcite .
the photoemission measurements were performed using a gammadata scienta ses2002 analyzer and monochromatic laboratory photon sources .
the energy resolutions were set to 0.4 ev , 5 mev and 5 mev at al @xmath13 ( 1486.6 ev ) , he ii@xmath14 ( 40.8 ev ) and he i@xmath14 ( 21.2 ev ) photon energies , respectively . the base pressure in the vacuum chamber was 3 @xmath15 10@xmath16 torr .
the temperature variation down to 20 k was achieved by an open cycle he cryostat , lt-3 m from advanced research systems , usa . the sample surface was cleaned by _ in situ _ scraping using a diamond file and the surface cleanliness was ensured by the absence of o 1@xmath17 and c 1@xmath17 features in the @xmath8-ray photoelectron ( xp ) spectra and the absence of impurity features in the binding energy range of 5 - 6 ev in the ultraviolet photoelectron ( up ) spectra .
the reproducibility of the spectra was confirmed after each trial of cleaning process .
the electronic band structure of ce@xmath0cosi@xmath1 was calculated using _ state - of - the - art _ full potential linearized augmented plane wave ( flapw ) method using wien2k software@xcite within the local density approximations , lda .
the convergence for different calculations were achieved considering 512 @xmath18 points within the first brillouin zone .
the error bar for the energy convergence was set to @xmath19 0.2 mev per formula unit ( fu ) . in every case , the charge convergence was achieved to be less than 10@xmath20 electronic charge .
the lattice constants used in these calculations are determined from the @xmath8-ray diffraction patterns considering alb@xmath0 derived hexagonal structure and are found to be @xmath21 = 8.104 and @xmath22 = 4.197 @xcite .
the muffin - tin radii ( @xmath23 ) for ce , co and si were set to 2.5 a.u .
, 2.28 a.u . and 2.02 a.u .
, respectively .
( xps ) , and ( b ) he ii and he i excitation energies .
the solid line in ( a ) represents the broadened he ii spectrum to take into account the energy resolution corresponding to xp spectrum .
dashed line represent the calculated xp spectrum as described later in the text .
( c ) the ce 4@xmath3 spectral function obtained by subtracting he i spectrum from the he ii spectrum . ] the valence band in ce@xmath0cosi@xmath1 consists of ce 5@xmath2 , ce 4@xmath3 , co 3@xmath2 and si 3@xmath5 electronic states . since the transition probability of the photoelectrons in the photo - excitation process strongly depends on the excitation energies , a comparison of the photoemission spectra collected at different excitation energies would help to identify experimentally the character of various features constituting the valence band . in fig .
1(a ) , we show the valence band spectra collected at 20 k using al k@xmath14 photon energy .
the he ii@xmath14 and he i@xmath14 spectra are shown in fig .
1(b ) . in order to compare the spectral functions , it is important to subtract suitable background from the raw data as shown in fig .
we have subtracted integral background from al k@xmath14 and he ii spectra .
since the thermalized electrons lead to a large increase in background at low kinetic energies , the background function in the he i spectrum are often found to be better defined by a polynomial ( quadratic like ) as shown in fig .
evidently , the background contributions in the energy range 0 - 3 ev binding energies is quite small in every case .
moreover , different functional dependence of the background in the he i spectrum does not have significant influence on the conclusions as the calculated spectra are compared with the xp spectrum .
excitation energies .
the open circles represent raw data , lines represent the background function and the solid circles are the subtracted spectra . ] in fig .
1(b ) , the intensity close to the fermi level , @xmath6 , exhibit significantly different behavior as a function of excitation energy .
he ii spectrum exhibits large intensity and two distinct features c and k near @xmath6 , which are not visible in the other spectra .
the high energy resolution employed in the he i and he ii measurements helped to resolve distinct signature of the features in the vicinity of @xmath6 .
the photoemission cross section for ce 4@xmath3 states at he ii excitation energy is about 3 times larger than that at he i photon energy while it is almost the same for si 3@xmath5 states and double for co 3@xmath2 states @xcite .
thus , the features c and k can be attributed to the photoemission signal primarily from the ce 4@xmath3 states .
we have subtracted the he i spectrum from the he ii spectrum to delineate the ce 4@xmath3 contributions . the subtracted spectrum representing ce 4@xmath3 band is shown in fig .
the distinct features , c and k corresponding to the spin orbit satellite of the abrikosov - suhl resonance ( asr ) and the main peak of asr , respectively , could clearly be identified @xcite .
in addition to the resonance feature at @xmath6 , each spectrum exhibits two distinct features , a and b at about 2.3 ev and 1 ev binding energies , respectively .
interestingly , the relative intensity of these features does not change with such a large change in photon energies .
this is an important observation as the photoemission cross sections corresponding to ce 5@xmath2 , ce 4@xmath3 , co 3@xmath2 and si 3@xmath5 electronic states have significantly different excitation energy dependence @xcite . to verify the change in lineshape with better clarity , we have broadened the he ii spectrum by convoluting a gaussian of full width at half maximum ( fwhm ) = 0.4 ev to make the resolution broadening comparable to that of the xp spectrum .
the broadened spectrum is shown by solid line superimposed over the xp spectrum in fig .
the lineshapes of both the spectra are almost identical . in fig .
1(b ) , the he i and he ii spectra are superimposed over each other to investigate the change in lineshape when the energy resolution broadening is the same ( @xmath11 5 mev ) .
the spectra in the binding energy range beyond 1 ev are found to be almost identical .
all these observations are unusual .
these observations can not be associated to the surface preparation via scraping for the following reasons . scraping helps to break the sample grains in an non - selective manner that helps to expose a clean surface for the measurements .
in addition , scraping usually enhances the surface roughness .
such a process will enhance the surface contribution in the spectra .
in addition , the linewidth of the features may also get enhanced due to scraping .
this later possibility can be ruled out as it is shown experimentally @xcite that cleaving and scraping have insignificant influence on the spectral lineshape at higher binding energies as the intrinsic linewidth of these features are already large due to various lifetime effects .
as for the enhancement of surface contributions , it is already known that photoemission with he i and he ii photons have significant surface sensitivity ( probing depth is @xmath11 8 - 10 ) .
thus , these spectra may contain large surface contributions due to scraping .
however , the al k@xmath14 valence band spectra possess dominant bulk contributions ( probing depth @xmath11 25 ) .
it is experimentally found that these spectra can often be considered as representative of the bulk electronic structure @xcite suggesting that the influence of scraping will have less influence in the al @xmath13 spectrum .
thus , the observations of similar lineshape in the energy range of 1 to 3 ev binding energies in figs . 1(a ) and 1(b ) suggest that either ( i ) the surface sensitivity in the he i and he ii spectra are compensated by scraping or ( ii ) the surface and bulk electronic structures are quite similar and scraping has insignificant influence in the spectral lineshape . the former is unlikely as scraping enhances the surface contribution that will make the differences more prominent .
thus , the reasonable conclusion will the later case .
partial density of states ( pdos ) , ( b ) ce 4@xmath3 pdos , ( c ) co 3@xmath2 pdos and ( d ) si 3@xmath5 pdos .
the thick solid line in ( b ) represents the ce 4@xmath3 pdos rescaled by 25 times to show the weak intensities at lower energies . ] in order to verify the character of the features theoretically , we have calculated the electronic band structure using fplapw method . the calculated partial density of states ( pdos )
are shown in fig .
the dominant contribution in this energy range arises from the ce 5@xmath2 , ce 4@xmath3 , co 3@xmath2 and si 3@xmath5 pdos as shown in fig .
3(a ) , 3(b ) , 3(c ) and 3(d ) respectively .
all the other contributions are negligible in this energy range . evidently , ce 5@xmath2 contributions are small and almost equally distributed over the whole energy range shown .
ce 4@xmath3 band is intense and narrow as expected . in order to provide clarity
, we have rescaled the ce 4@xmath3 partial density of states ( pdos ) by 25 times and shown by thick solid line in fig .
the 4@xmath3 pdos contribute essentially in the vicinity of the fermi level ( @xmath19 1 ev binding energy ) .
the intensity of ce 4@xmath3 band is significantly weak at higher binding energies .
this is consistent with the observation in fig .
co 3@xmath2 states also have finite contributions in this energy range due to the hybridization between co 3@xmath2 and ce 4@xmath3 states .
co 3@xmath2 and si 3@xmath5 electronic states are strongly hybridized and appear dominantly in the binding energy range larger than 0.5 ev .
the bonding states contribute in the energy range higher than 2 ev binding energy , where the si 3@xmath5 pdos has large contributions .
the antibonding features appear in the energy range 0.5 to 2 ev , where co 3@xmath2 contributions are dominant .
this suggests that the feature b in fig .
1 has dominant co 3@xmath2 character and the intensities corresponding to si 3@xmath5 photoemission contribute to feature a. the dominance of co 3@xmath2 contributions in this whole energy range shown presumably leads to unchanged spectral lineshape with the change in photon energy as observed in fig 1(a ) and 1(b ) . in order to compare the experimental spectrum with the calculated results
, we have calculated the xp spectrum in the following way : the ce 5@xmath2 , ce 4@xmath3 , co 3@xmath2 and si 3@xmath5 pdos per formula unit were multiplied by the corresponding photoemission cross sections at al @xmath13 energy .
the sum of all these contributions was convoluted by the fermi distribution function at 20 k and broadened by the lorentzian function to account for the photo - hole lifetime broadening .
the resolution broadening is introduced via further broadening of the spectrum by a gaussian function of fwhm = 0.4 ev .
the calculated spectrum is shown by dashed line in fig .
1(a ) after normalizing by the total integrated area under the curve .
the intensities near @xmath6 in the experimental spectrum appears to be captured reasonably well in the calculated spectra . in the higher binding energy region ,
the intensity around 1 ev is overestimated and that around 2.5 ev is underestimated ( see fig .
since this energy range contains dominant contribution from the co 3@xmath2 states , it naturally indicates that co 3@xmath2 pdos region is not well described and correlations among co 3@xmath2 electrons may be important in determining the electronic structure in this energy range . in order to verify this
, we have calculated the electronic density of states considering finite electron correlation , @xmath24 , among co 3@xmath2 electrons .
the spectral functions corresponding to different @xmath24 values are calculated from the lda+@xmath4 results following the procedure described above
. -values .
( a ) the calculated spectral functions without ce 4@xmath3 contributions .
( b ) the calculated spectral functions contains ce 4@xmath3 contributions .
clearly , the results in ( b ) provide better description than that in ( a ) revealing signature of ce 4@xmath3 contributions in the vicinity of the fermi level . ] the lines in fig .
4 represent the spectral functions for different @xmath24 values which are superimposed on the experimental xp spectrum represented by open circles . in fig .
4(a ) we show the calculated spectra without ce 4@xmath3 contributions and the ones including ce 4@xmath3 contributions are shown in fig .
it is evident from the figure that the signature of the feature around 2.3 ev ( feature a ) becomes more and more prominent with the increase in @xmath24 .
consequently , the intensity of feature b reduces .
thus , the feature a can be attributed to the photoemission signal from electron correlation induced co 3@xmath2 bands ( lower hubbard band ) in addition to the si 3@xmath5 contributions .
it is to note here that although the feature a has comparable contributions from co 3@xmath2 and si 3@xmath5 states from band structure calculations , the photoemission intensity from the co 3@xmath2 states would be larger than that of si 3@xmath5 states at xps photon energies due to the matrix element effects .
it is clear that the calculated spectral functions corresponding to @xmath24 @xmath11 3 ev is close to the experimental spectrum compared to all other cases . for higher values of @xmath24 ,
the correlation induced feature becomes stronger and appears at higher binding energies .
the finding of such correlation among co 3@xmath2 electrons is not unrealistic as it is often found in the literature that co 3@xmath2 electrons are quite strongly correlated @xcite .
it is to note here that the estimation of @xmath24 obtained within the lda+@xmath4 calculations in this study suggests signature of significant correlation among co 3@xmath2 electrons .
an accurate estimate of @xmath24 requires further study using other methods e.g. lda+dmft .
partial density of states for different electron correlation strength , @xmath25 among ce 4@xmath3 electrons .
the results corresponding to the whole energy range is shown in ( a ) and that near fermi level is shown in ( b ) . ]
the comparison of fig .
4(a ) and 4(b ) establishes that the intensities near @xmath6 essentially arise due to the photoemission from the occupied part of the ce 4@xmath3 band . the electron correlation among the co
3@xmath2 electrons has negligible influence on the spectral intensity at @xmath6 .
consideration of ce 4@xmath3 bands in the spectral function calculation leads to a better description of the spectral intensities at the fermi level .
electron correlation among ce 4@xmath3 electrons is also known to be strong .
thus , we verify if such an effect leads to significant spectral intensity at higher binding energies .
we show the ce 4@xmath3 pdos calculated for @xmath25 = 0 and 4 ev in fig .
the unoccupied part of the spectral function exhibits large spectral redistribution .
the occupied part exhibits a change in lineshape close to the fermi level as shown with better clarity in fig .
evidently , the 4@xmath3 contributions appear essentially in the vicinity of the fermi level even if the 4@xmath3 correlations are considered .
spectrum ( open circles ) exhibiting distinct signature of main and satellite peaks .
the satellite intensity is shown by shaded region .
solid line passing through the experimental data represents the fit comprising of a main peak ( dashed line ) and a satellite ( shaded peak ) . ]
the signature of electron correlations can also be observed in the co 2@xmath5 core level spectra @xcite . in fig .
6 , we show the co 2@xmath5 spectrum collected at 20 k using al @xmath13 radiation .
the spectrum consists of two spin orbit split features co 2@xmath26 and co 2@xmath27 at 778 ev and 792.9 ev binding energies , respectively ( energy separation of about 14.9 ev ) .
these binding energies are identical to those found in elemental co metals @xcite .
this indicates that the valence state of co in this material is very similar to that of the elemental co metals .
in addition , every spin orbit split feature exhibits a weak but distinct shoulder at higher binding energies .
this has been shown more clearly by rescaling and shifting this energy region in the same figure .
although the satellite intensity looks small , it appears to be quite intense compared to the ones observed in earlier studies @xcite .
the energy separation between the main peak and the satellite is about 4 ev .
it is well known that the core level spectra reveal multiple features representing differently screened final states .
the intense feature at lower binding energy is usually called the well screened feature and the ones at higher binding energies are called poorly screened features / satellites .
such satellites appear due to the finite electron correlations .
the finite intensity of the satellite feature in the present case , independently establishes the presence of electronic correlations among co 3@xmath2 electrons as concluded from valence band spectra @xcite .
thus , our results establish that the electron correlation strength among co 3@xmath2 electrons are significant and needs consideration to derive the electronic structure of these systems .
the mixed valency and/or kondo effect in this class of compounds depends on the kondo coupling strength , @xmath28 , where @xmath29 is the hybridization strength between conduction and the 4@xmath3 electronic states , @xmath30 is the binding energy of the 4@xmath3 electrons . in the present case ,
co 3@xmath2 electrons contribute significantly in the formation of the conduction band .
electron correlation introduces local character among 3@xmath2 electronic states ; the band dispersion will be weaker than the case without electron correlation and hence the hybridization strength will be weaker .
in addition , the formation of the lower hubbard band at higher binding energies relative to the uncorrelated band due to the electron correlation will influence the effective energy separation between the ce 4@xmath3 band and the conduction electrons .
thus , @xmath31 and various hopping parameters involving co 3@xmath2 electronic states are expected to reduce leading to an effective reduction in the degree of mixed valency of ce .
a quantitative estimation of such effects requires detailed modeling of this case presumably within the periodic anderson model .
we believe that these results will help to initiate such research by the theoreticians .
in summary , we have studied the electronic structure of ce@xmath0cosi@xmath1 using high resolution photoemission spectroscopy and _ ab initio _ band structure calculations .
the experimental results indicate the dominance of co 3@xmath2 contributions in the valence band .
si 3@xmath5 states appear at higher binding energies ( 2 - 3 ev ) .
the ce 4@xmath3 contributions appear essentially in the vicinity of the fermi level .
high resolution employed in this study helped to probe the kondo resonance feature appearing at the fermi level .
although the contribution of the co 3@xmath2 states at the fermi level is weak , co 3@xmath2 states are found to be hybridized with the ce 4@xmath3 states .
the comparison of the experimental results with the calculated ones reveal distinct signature of electron correlation among co 3@xmath2 electrons .
these results suggests that description of the electronic structure and various interesting electronic properties involving kondo systems requires consideration of correlation among conduction electrons .
one of the authors s.p . , thanks the council of scientific and industrial research , government of india for financial support .
brandt n b and moshchalkov v v 1984 _ advances in physics _ * 33 * 373 wills john m and cooper bernard r 1987 _ phys .
b _ * 36 * 3809 gunnarsson o and jepsen o 1988 _ phys . rev .
b _ * 38 * 3568 gunnarsson o , andersen o k , jepsen o and zaanen j 1989 _ phys . rev . b _ * 39 * 1708 gunnarsson o and schnhammer k 1989 _ phys . rev .
b _ * 40 * 4160 gordon r a , alexander m g , warren c j , disalvo f j and pttgen r 1997 _ j. alloys compounds _ * 248 * 24 majumdar subham , mahesh kumar m , mallik r and sampathkumaran e v 1999 _ solid state comm . _ * 110 * 509 patil swapnil , iyer kartik k , maiti k and sampathkumaran e v 2008 _ phys .
b _ * 77 * 094443 ; patil swapnil , iyer kartik k , maiti k and sampathkumaran e v condmat-0803.0652 lawrence j m , riseborough p s and parks r d 1981 _ rep . prog .
_ * 44 * 1 blaha p , schwarz k , madsen g k h , kvasnicka d and luitz j 2001 * wien2k * , an augmented plane wave + local orbitals program for calculating crystal properties ( karlheinz schwarz , techn .
universitt wien , austria ) , isbn 3 - 9501031 - 1 - 2 yeh j j and lindau i 1985 _ at .
data nucl .
data tables _ * 32 * , 1 maiti k , manju u , ray s , mahadevan p , inoue i h , carbone c and sarma d d 2006 _ phys .
b _ * 73 * 052508 maiti k , kumar a , sarma d d , weschke e and kaindl g 2004 _ phys .
b _ * 70 * 195112 maiti k and sarma d d , 2000 _ phys . rev .
b _ * 61 * 2525 maiti k and singh r s 2005 _ phys . rev .
b _ * 71 * 161102(r ) sekiyama a and suga s 2002 _ physica b _ * 312 - 313 * 634 ehm d , hfner s , reinert f , kroha j , wlfle p , stockert o , geibel c and lhneysen h v 2007 _ phys . rev .
b _ * 76 * 045117 chainani a , mathew m and sarma d d 1992 _ phys . rev .
b _ * 46 * , 9976 wu z y , benfatto m , pedio m , cimino r , mobilio s , barman s r , maiti k and sarma d d 1997 _ phys . rev .
b _ * 56 * 2228 saitoh t , mizokawa t , fujimori a , abbate m , takeda y and takano m 1997 _ phys .
b _ * 55 * 4257 pandey s k , kumar a , patil s , medicherla v r r , singh r s , maiti k , prabhakaran d , boothroyd a t and pimpale a v 2008 _ phys .
b _ * 77 * 045123 pandey s k , patil s , medicherla v r r , singh r s and maiti k 2008 _ phys . rev . b _ * 77 * 115137 imada masatoshi , fujimori atsushi and tokura yoshinori 1998 _ rev . mod . phys . _ * 70 * 1039 nath krishna g , haruyama y and kinoshita t 2001 _ phys . rev .
b _ * 64 * 245417 schneider c m , pracht u , kuch w , chass a and kirschner j 1996 _ phys . rev .
b _ * 54 * r15618 | kondo systems are usually described by the interaction of strong correlation induced local moment with the highly itinerant conduction electrons . here , we study the role of electron correlations among conduction electrons in the electronic structure of a kondo lattice compound , ce@xmath0cosi@xmath1 , using high resolution photoemission spectroscopy and _ ab initio _ band structure calculations , where co 3@xmath2 electrons contribute in the conduction band .
high energy resolution employed in the measurements helped to reveal signature of ce 4@xmath3 states derived kondo resonance feature at the fermi level and dominance of co 3@xmath2 contributions at higher binding energies in the conduction band .
the line shape of the experimental co 3@xmath2 band is found to be significantly different from that obtained from the band structure calculations within the local density approximations , lda .
consideration of electron - electron coulomb repulsion , @xmath4 among co 3@xmath2 electrons within the lda+@xmath4 method leads to a better representation of experimental results .
signature of electron correlation induced satellite feature is also observed in the co 2@xmath5 core level spectrum .
these results clearly demonstrate the importance of the electron correlation among conduction electrons in deriving the microscopic description of such kondo systems . |
a statistical mixture model @xcite @xmath1 with @xmath2 weighted components has underlying probability distribution : @xmath3 with @xmath4 and @xmath5 denoting the mixture parameters : the @xmath6 s are positive weights summing up to one , and the @xmath7 s denote the individual component parameters . (
appendix [ sec : notations ] summarizes the notations used throughout the paper . )
mixture models of @xmath8-dimensional gaussians are the most often used statistical mixtures @xcite . in that case
, each component distribution @xmath9 is parameterized by a mean vector @xmath10 and a covariance matrix @xmath11 that is symmetric and positive definite .
that is , @xmath12 .
the gaussian distribution has the following probability density defined on the support @xmath13 : @xmath14 where @xmath15 denotes the squared mahalanobis distance @xcite @xmath16 defined for a symmetric positive definite matrix @xmath17 ( @xmath18 , the precision matrix ) . to draw a random variate from a gaussian mixture model ( gmm ) with @xmath0 components ,
we first draw a multinomial variate @xmath19 , and then sample a gaussian variate from @xmath20 . a multivariate normal variate @xmath21
is drawn from the chosen component @xmath22 as follows : first , we consider the cholesky decomposition of the covariance matrix : @xmath23 , and take a @xmath8-dimensional vector with coordinates being random standard normal variates : @xmath24^t$ ] with @xmath25 ( for @xmath26 and @xmath27 uniform random variates in @xmath28 ) .
finally , we assemble the gaussian variate @xmath21 as @xmath29 .
this drawing process emphasizes that sampling a statistical mixture is a _ doubly stochastic process _ by essence : first , we sample a multinomial law for choosing the component , and then we sample the variate from the selected component .
figure [ fig : gmm5d](b ) shows a gmm with @xmath30 components learned from a color image modeled as a 5d xyrgb point set ( figure [ fig : gmm5d](a ) ) . since
a gmm is a _ generative model _ , we can sample the gmm to create a `` sample image '' as shown in figure [ fig : gmm5d](c ) .
observe that low frequency information of the image is nicely modeled by gmms .
figure [ fig : gmmhighd](f ) shows a gmm with @xmath30 components learned from a color image modeled as a high - dimensional point set .
each @xmath31 color image patch anchored at @xmath32 is modeled as a point in dimension @xmath33 .
gmm representations of images and videos @xcite provide a compact feature representation that can be used in many applications , like in information retrieval ( ir ) engines @xcite .
( a ) components is trained ( b ) .
drawing many random variates from the generative gmm yields a sample image(c ) that keeps low - frequency visual information .
[ fig : gmm5d],title="fig:",scaledwidth=25.0% ] ( b ) components is trained ( b ) .
drawing many random variates from the generative gmm yields a sample image(c ) that keeps low - frequency visual information .
[ fig : gmm5d],title="fig:",scaledwidth=25.0%,scaledwidth=25.0% ] ( c ) components is trained ( b ) .
drawing many random variates from the generative gmm yields a sample image(c ) that keeps low - frequency visual information .
[ fig : gmm5d],title="fig:",scaledwidth=25.0% ] in this paper , we consider the general case of mixtures of distributions belonging the same exponential family @xcite , like gaussian mixture models @xcite ( gmms ) , rayleigh mixture models @xcite ( rmms ) , laplacian mixture models ( lmms)@xcite , bernoulli mixture models @xcite ( bmms ) , multinomial mixture models @xcite ( mmms ) , poisson mixture models ( pmms ) @xcite , weibull mixture models @xcite ( weimms ) , wishart mixture models @xcite ( wismm ) , etc .
ccc -gmm modeling depicted by its covariance ellipses ,
( c ) hard segmentation using the gmm , ( d ) sampling the 5d gmm , ( e ) mean colors ( @xmath34 patches ) for gmm with patch size @xmath35 , ( f ) patch mean @xmath36 for @xmath35 patch size width.[fig : gmmhighd],title="fig : " ] & -gmm modeling depicted by its covariance ellipses , ( c ) hard segmentation using the gmm , ( d ) sampling the 5d gmm , ( e ) mean colors ( @xmath34 patches ) for gmm with patch size @xmath35 , ( f ) patch mean @xmath36 for @xmath35 patch size width.[fig : gmmhighd],title="fig : " ] & -gmm modeling depicted by its covariance ellipses , ( c ) hard segmentation using the gmm , ( d ) sampling the 5d gmm , ( e ) mean colors ( @xmath34 patches ) for gmm with patch size @xmath35 , ( f ) patch mean @xmath36 for @xmath35 patch size width.[fig : gmmhighd],title="fig : " ] + ( a ) & ( b ) & ( c ) + -gmm modeling depicted by its covariance ellipses , ( c ) hard segmentation using the gmm , ( d ) sampling the 5d gmm , ( e ) mean colors ( @xmath34 patches ) for gmm with patch size @xmath35 , ( f ) patch mean @xmath36 for @xmath35 patch size width.[fig : gmmhighd],title="fig : " ] & -gmm modeling depicted by its covariance ellipses , ( c ) hard segmentation using the gmm , ( d ) sampling the 5d gmm , ( e ) mean colors ( @xmath34 patches ) for gmm with patch size @xmath35 , ( f ) patch mean @xmath36 for @xmath35 patch size width.[fig : gmmhighd],title="fig : " ] & -gmm modeling depicted by its covariance ellipses , ( c ) hard segmentation using the gmm , ( d ) sampling the 5d gmm , ( e ) mean colors ( @xmath34 patches ) for gmm with patch size @xmath35 , ( f ) patch mean @xmath36 for @xmath35 patch size width.[fig : gmmhighd],title="fig : " ] + + ( d ) & ( e ) & ( f ) expectation - maximization @xcite ( em ) is a traditional algorithm for learning finite mixtures @xcite . banerjee et al .
@xcite proved that em for mixture of exponential families amounts to perform equivalently a soft bregman clustering .
furthermore , this em - bregman soft clustering equivalence was extended to total bregman soft clustering for curved exponential families @xcite .
although mathematically convenient , we should remember that mixture data should be hard clustered as each observation should emanate from exactly one component .
it is well - known that @xmath0-means clustering technique can be interpreted as a limit case of em for isotropic gaussian mixtures @xcite .
kearns et al .
@xcite casted further light on the hard / soft relationship using an information - theoretic analysis of hard @xmath0-means and soft expectation - mazimization assignments in clustering .
banerjee et al @xcite proved a mathematical equivalence between the estimation of maximum likelihood of exponential family mixtures ( mlme , maximum likelihood mixture estimation ) and a rate distortion problem for bregman divergences
. furthermore , banerjee et al .
@xcite proposed the hardened expectation for the special case of von mises - fisher mixtures ( hard em , section 4.2 of @xcite ) for computational efficiency . in this paper
, we build on the duality between bregman divergences and exponential families @xcite to design @xmath0-mle that iteratively ( 1 ) assigns data to mixture components , ( 2 ) update mixture parameters la @xmath0-means and repeat step ( 1 ) until local convergence , ( 3 ) update weights and reiterate from ( 1 ) until local convergence ( see algorithm [ algo : kmle ] ) .
we prove that @xmath0-mle maximizes monotonically the complete likelihood function .
we also discuss several initialization strategies and describe a probabilistic initialization @xmath0-mle++ with guaranteed performance bounds .
the paper is organized as follows : section [ sec : preliminaries ] recall the basic notions of exponential families , legendre transform , bregman divergences , and demonstrate the duality between bregman divergences and exponential families to study the maximum likelihood estimator ( mle ) . section [ sec : kmle - theta ] presents the framework of @xmath0-mle for mixtures with prescribed weights , based on the bregman - exponential family duality .
the generic @xmath0-mle algorithm is described in section [ sec : fullkmle ] , and section [ sec : speedup ] discusses on proximity location data - structures to speed up the assignment step of the algorithm .
section [ sec : kmlepp ] presents @xmath0-mle++ , a probabilistic initialization of @xmath0-mle .
finally , section [ sec : concl ] concludes the paper and discusses on avenues for future research .
an exponential family @xcite @xmath37 is a set of parametric probability distributions @xmath38 whose probability density .
we do not introduce the framework of probability measures nor radon - nikodym densities .
] can be decomposed canonically as @xmath39 where @xmath40 denotes the sufficient statistics , @xmath41 the natural parameter , @xmath42 the log - normalizer , and @xmath43 a term related to an optional auxiliary carrier measure .
@xmath44 denotes the inner product ( i.e. , @xmath45 for vectors @xmath46 for matrices , etc . ) .
let @xmath47 denotes the natural parameter space .
the dimension @xmath48 of the natural parameter space is called the order of the family . for the @xmath8-variate gaussian distribution ,
the order is @xmath49 .
it can be proved using the cauchy - schwarz inequality @xcite that the log - normalizer @xmath50 is a strictly convex and differentiable function on an open convex set @xmath51 .
the log - density of an exponential family is @xmath52 to build an exponential family , we need to choose a basic density measure on a support @xmath53 , a sufficient statistic @xmath40 , and an auxiliary carrier measure term @xmath43 . taking the log - laplace transform , we get @xmath54 and define the natural parameter space as the @xmath41 values ensuring convergence of the integral . in fact , many usual statistical distributions such as the gaussian , gamma , beta , dirichlet , poisson , multinomial , bernoulli , von mises - fisher , wishart , weibull are exponential families in disguise . in that case , we start from their probability density or mass function to retrieve the canonical decomposition of eq .
[ eq : cexpfam ] .
see @xcite for usual canonical decomposition examples of some distributions that includes a bijective conversion function @xmath55 for going from the usual @xmath56-parameterization of the distribution to the @xmath41-parametrization .
furthermore , exponential families can be parameterized canonically either using the natural coordinate system @xmath41 , or by using the dual moment parameterization @xmath57 ( also called mean value parameterization ) arising from the legendre transform ( see appendix [ sec : mvn ] for the case of gaussians ) . for a strictly convex and differentiable function @xmath58
, we define its convex conjugate by @xmath59 the maximum is obtained for @xmath60 and is unique since @xmath50 is convex @xmath61 : @xmath62 thus strictly convex and differentiable functions come in pairs @xmath63 with gradients being functional inverses of each other @xmath64 and @xmath65 .
legendre transform is an involution : @xmath66 for strictly convex and differentiable functions . in order to compute @xmath67
, we only need to find the functional inverse @xmath68 of @xmath69 since @xmath70 however , this inversion may require numerical solving when no analytical expression of @xmath71 is available . see for example the gradient of the log - normalizer of the gamma distribution @xcite , the dirichlet or von mises - fisher distributions @xcite . a bregman divergence
@xmath72 is defined for a strictly convex and differentiable generator @xmath50 as @xmath73 the kullback - leibler divergence ( relative entropy ) between two members @xmath74 and @xmath75 of the same exponential family amounts to compute a bregman divergence on the corresponding swapped natural parameters : @xmath76 the proof follows from the fact that @xmath77=\int_{x\in\mathbb{x } } t(x)p_f(x;\theta)\dx=\nabla f(\theta)$ ] @xcite . using legendre transform , we further have the following equivalences of the relative entropy : @xmath78 where @xmath60 is the dual moment parameter ( and @xmath79 ) .
information geometry @xcite often considers the canonical divergence @xmath80 of eq .
[ eq : cd ] that uses the mixed coordinate systems @xmath81 , while computational geometry @xcite tends to consider dual bregman divergences , @xmath72 or @xmath82 , and visualize structures in one of those two canonical coordinate systems .
those canonical coordinate systems are dually orthogonal since @xmath83 , the identity matrix . for exponential family mixtures with a single component @xmath84 ( @xmath85 , @xmath86 )
, we easily estimate the parameter @xmath87 . given @xmath88 independent and identically distributed observations @xmath89 , the maximum likelihood estimator ( mle ) is maximizing the likelihood function : @xmath90 for exponential families , the mle reports a unique maximum since the hessian of @xmath50 is positive definite ( @xmath91\succ 0 $ ] ) : @xmath92 the mle is consistent and efficient with asymptotic normal distribution : @xmath93 where @xmath94 denotes the fisher information matrix : @xmath95 = \nabla ^2 f(\theta ) = ( \nabla^2 g(\eta))^{-1}\ ] ] ( this proves the convexity of @xmath50 since the covariance matrix is necessarily positive definite . ) note that the mle may be biased ( for example , normal distributions ) . by using the legendre transform ,
the log - density of an exponential family can be interpreted as a bregman divergence @xcite : @xmath96 table [ tab : duality ] reports some illustrating examples of the bregman divergence @xmath97 exponential family duality .
.some examples illustrating the duality between exponential families and bregman divergences.[tab : duality ] [ cols="^,^,^ " , ] * \0 . *
initialization * : * * calculate global mean @xmath98 and global covariance matrix @xmath99 : @xmath100 * * @xmath101 , initialize the @xmath102th seed as @xmath103
* assignment * : + @xmath104 with @xmath105 the squared mahalanobis distance : @xmath106 .
+ let @xmath107 be the cluster partition : @xmath108 .
+ ( anisotropic voronoi diagram @xcite ) * \2 . *
update the parameters * : + @xmath109 + * goto step 1 * unless local convergence of the complete likelihood is reached .
update the mixture weights * : @xmath110 .
+ * goto step 1 * unless local convergence of the complete likelihood is reached .
the @xmath0-mle++ initialization for the gmm is reported in algorithm [ algo : kmleppgmm ] .
* choose first seed @xmath111 , for @xmath112 uniformly random in @xmath113 .
* for @xmath114 to @xmath0 * * choose @xmath115 with probability + @xmath116 where @xmath117 .
+ @xmath118 * * add selected seed to the initialization seed set : @xmath119 .
we instantiate the soft bregman em , hard em , @xmath0-mle , and @xmath0-mle++ for the rayleigh distributions , a sub - family of weibull distributions .
a rayleigh distribution has probability density @xmath120 where @xmath121 denotes the _ mode _ of the distribution , and @xmath122 the support .
the rayleigh distributions form a @xmath123-order univariate exponential family ( @xmath124 ) .
re - writing the density in the canonical form @xmath125 , we deduce that @xmath126 , @xmath127 , @xmath128 , and @xmath129 .
thus @xmath130 and @xmath131 .
the natural parameter space is @xmath132 and the moment parameter space is @xmath133 ( with @xmath134 ) .
we check that conjugate gradients are reciprocal of each other since @xmath135 , and we have @xmath136 ( i.e , dually orthogonal coordinate system ) with @xmath137 and @xmath138 .
rayleigh mixtures are often used in ultrasound imageries @xcite . following banerjee et al .
@xcite , we instantiate the bregman soft clustering for the convex conjugate @xmath139 , @xmath126 and @xmath134 . the rayleigh density expressed in the @xmath57-parameterization yields @xmath140
. expectation .
: : soft membership for all observations @xmath89 : + @xmath141 + ( we can use any of the equivalent @xmath142 , @xmath41 or @xmath57 parameterizations for calculating the densities . ) maximization .
: : barycenter in the moment parameterization : + @xmath143 the associated bregman divergence for the convex conjugate generator of the rayleigh distribution log - normalizer is @xmath144 this is the itakura - saito divergence is ( indeed , @xmath67 is equivalent modulo affine terms to @xmath145 , the burg entropy ) . 1 . hard assignment .
: : @xmath146 + voronoi partition into clusters : + @xmath147 2 .
@xmath57-parameter update .
: : @xmath148 + @xmath149 + go to 1 . until ( local ) convergence is met .
weight update .
: : @xmath150 + go to 1 . until ( local ) convergence is met .
note that @xmath0-mle does also model selection as it may decrease the number of clusters in order to improve the complete log - likelihood .
if initialization is performed using random point and uniform weighting , the first iteration ensures that all voronoi cells are non - empty .
a good initialization for rayleigh mixture models is done as follows : compute the order statistics for the @xmath151-th elements ( in overall @xmath152-time ) .
those pivot elements split the set @xmath53 into @xmath0 groups @xmath153 of size @xmath154 , on which we estimate the mles .
the @xmath0-mle++ initialization is built from the itakura - saito divergence : @xmath155 k - mle++ : * choose first seed @xmath111 , for @xmath112 uniformly random in @xmath113 . * for @xmath114 to @xmath0 * * choose @xmath156 with probability + @xmath157 * * add selected seed to the initialization seed set : @xmath119 .
ll : + @xmath44 & inner product ( e.g. , @xmath158 for vectors , @xmath159 for matrices ) + @xmath160 & exponential distribution parameterized using the @xmath41-coordinate system + @xmath161 & support of the distribution family ( @xmath162 ) + @xmath8 & dimension of the support @xmath161 ( univariate versus multivariate ) + @xmath48 & dimension of the natural parameter space + & ( uniparameter versus multiparameter ) + @xmath40 & sufficient statistic ( @xmath163 ) + @xmath43 & auxiliary carrier term + @xmath50 & log - normalizer , log - laplace , cumulant function ( @xmath164 ) + @xmath69 & gradient of the log - normalizer ( for moment @xmath57-parameterization ) + @xmath165 & hessian of the log - normalizer + & ( fisher information matrix , spd : @xmath166 ) + @xmath67 & legendre convex conjugate + : + @xmath41 & canonical natural parameter + @xmath167 & natural parameter space + @xmath57 & canonical moment parameter + @xmath168 & moment parameter space + @xmath56 & usual parameter + @xmath169 & usual parameter space + @xmath170 & density or mass function using the usual @xmath56-parameterization + @xmath171 & density or mass function using the usual moment parameterization + : + @xmath172 & mixture model + @xmath173 & closed probability @xmath174-dimensional simplex + @xmath175 & shannon entropy @xmath176 ( with @xmath177 by convention ) + @xmath178 & shannon cross - entropy @xmath179 + @xmath6 & mixture weights ( positive such that @xmath180 ) + @xmath7 & mixture component natural parameters + @xmath181 & mixture component moment parameters + @xmath182 & estimated mixture + @xmath0 & number of mixture components + @xmath183 & mixture parameters + : + @xmath184 & sample ( observation ) set + @xmath185 & cardinality of sets : @xmath88 for the observations , @xmath0 for the cluster centers + @xmath186 & hidden component labels + @xmath187 & sample sufficient statistic set + @xmath188 & likelihood function + @xmath189 & maximum likelihood estimates + @xmath190 & soft weight for @xmath191 in cluster / component @xmath192 ( @xmath193 ) + @xmath102 & index on the sample set @xmath194 + @xmath195 & index on the mixture parameter set @xmath196 + @xmath197 & cluster partition + @xmath198 & cluster centers + @xmath199 & cluster proportion size + @xmath72 & bregman divergence with generator @xmath50 : + & + @xmath201 & jensen diversity index : + & @xmath202 + : + @xmath203 & average incomplete log - likelihood : + & @xmath204 + @xmath205 & average complete log - likelihood + & @xmath206 + @xmath207 & geometric average incomplete likelihood : + & @xmath208 + @xmath209 & geometric average complete likelihood : + & @xmath210 + @xmath211 & average @xmath0-means loss function ( average divergence to the closest center ) + & cdric archambeau , john aldo lee , and michel verleysen . on convergence problems of the em algorithm for finite gaussian mixtures . in _
european symposium on artificial neural networks ( esann ) _ , pages 99106 , 2003 .
arindam banerjee , inderjit dhillon , joydeep ghosh , and srujana merugu .
an information theoretic analysis of maximum likelihood mixture estimation for exponential families . in _ proceedings of the twenty - first international conference on machine learning _ , icml , pages 5764 , new york , ny , usa , 2004 .
acm .
jason v. davis and inderjit s. dhillon .
differential entropic clustering of multivariate gaussians . in bernhard scholkopf ,
john platt , and thomas hoffman , editors , _ neural information processing systems ( nips ) _ , pages 337344 . mit press , 2006 .
kittipat kampa , erion hasanbelliu , and jose principe .
closed - form cauchy - schwarz pdf divergence for mixture of gaussians . in _ proceeding of the international joint conference on neural networks ( ijcnn ) _ , pages 2578
2585 , 2011 .
michael kearns , yishay mansour , and andrew y. ng .
an information - theoretic analysis of hard and soft assignment methods for clustering . in
_ proceedings of the thirteenth conference on uncertainty in artificial intelligence _ ,
uai , pages 282293 , 1997 .
franois labelle and jonathan richard shewchuk .
anisotropic voronoi diagrams and guaranteed - quality anisotropic mesh generation . in _ proceedings of the nineteenth annual symposium on computational geometry _ , scg 03 , pages 191200 , new york , ny , usa , 2003 .
acm .
james b. macqueen .
some methods of classification and analysis of multivariate observations . in l.
m. le cam and j. neyman , editors , _ proceedings of the fifth berkeley symposium on mathematical statistics and probability_. university of california press , berkeley , ca , usa , 1967 .
frank nielsen , paolo piro , and michel barlaud .
bregman vantage point trees for efficient nearest neighbor queries . in _
ieee international conference on multimedia and expo ( icme ) _ , pages 878881 , new york city , usa , june 2009 . ieee .
richard nock , panu luosto , and jyrki kivinen .
mixed bregman clustering with approximation guarantees . in _ proceedings of the european conference on machine learning and knowledge discovery in databases _ , pages 154169 , berlin , heidelberg , 2008 .
springer - verlag .
paolo piro , frank nielsen , and michel barlaud .
tailored bregman ball trees for effective nearest neighbors . in _
european workshop on computational geometry ( eurocg ) _ , loria , nancy , france , march 2009 . | we describe @xmath0-mle , a fast and efficient local search algorithm for learning finite statistical mixtures of exponential families such as gaussian mixture models .
mixture models are traditionally learned using the expectation - maximization ( em ) soft clustering technique that monotonically increases the incomplete ( expected complete ) likelihood . given prescribed mixture weights , the hard clustering @xmath0-mle algorithm iteratively assigns data to the most likely weighted component and update the component models using maximum likelihood estimators ( mles ) . using the duality between exponential families and bregman divergences ,
we prove that the local convergence of the complete likelihood of @xmath0-mle follows directly from the convergence of a dual additively weighted bregman hard clustering .
the inner loop of @xmath0-mle can be implemented using any @xmath0-means heuristic like the celebrated lloyd s batched or hartigan s greedy swap updates .
we then show how to update the mixture weights by minimizing a cross - entropy criterion that implies to update weights by taking the relative proportion of cluster points , and reiterate the mixture parameter update and mixture weight update processes until convergence .
hard em is interpreted as a special case of @xmath0-mle when both the component update and the weight update are performed successively in the inner loop . to initialize @xmath0-mle ,
we propose @xmath0-mle++ , a careful initialization of @xmath0-mle guaranteeing probabilistically a global bound on the best possible complete likelihood .
# 1#2#1,#2 # 1#2#1,#2 # 1doi:#1 exponential families , mixtures , bregman divergences , expectation - maximization ( em ) , @xmath0-means loss function , lloyd s @xmath0-means , hartigan and wong s @xmath0-means , hard em , sparse em . |
the solar coronal magnetic field is formed by active regions emerging from the solar interior and becoming distributed throughout the atmosphere by flux transport processes ( e.g. , mackay and yeates 2012 ) .
the field can become open to the heliosphere in regions where a single magnetic polarity dominates , including the polar caps and other regions of isolated polarity ( harvey and recely 2002 ) .
wherever the coronal fields extend to great heights , the magnetic tension force can be overcome by the thermal pressure of the expanding coronal plasma , and the field dragged out and opened to form a coronal hole ( parker 1958 ) . in extreme ultraviolet ( euv ) and soft x - rays ( sxr ) coronal holes appear as dark regions whereas in helium images they appear bright .
harvey and recely ( 2002 ) identified three classes of coronal hole : polar coronal holes confined to high latitudes ( @xmath0 , @xmath1 ) ; isolated coronal holes at active latitudes associated with the remnant fields of decayed active regions ; and transient coronal holes that briefly form after coronal mass ejections . from a study of the polar coronal holes during cycles 22 and 23 ,
harvey and recely concluded that polar coronal holes are formed from isolated holes associated with decayed active - region field of trailing polarity .
earlier , harvey and sheeley ( 1979 ) had shown that during the ascending phase of a solar cycle the leading polarities of active regions may diffuse to form locally unbalanced flux patterns that tend to form coronal holes ahead of active regions . the declining phase of cycle 23 roughly coincided with the beginning of nso s global oscillations network group ( gong ) synoptic magnetogram program and the launch of nasa s solar terrestrial relations observatory ( stereo ) mission ( kaiser et al .
2008 ) in late 2006 .
the cycle 23 minimum produced unusually many coronal holes at active latitudes compared to previous cycle minima , mainly owing to the weakness of the polar fields during cycle 23 ( de toma 2011 ) .
the decline of solar cycle 23 and the ascent of cycle 24 have provided excellent conditions for studying isolated coronal holes . using mt .
wilson observatory and wilcox solar observatory synoptic magnetograms and stereo euv synoptic maps , wang et al .
( 2010 ) studied three instances of small coronal holes appearing at the edges of newly emerged active regions and then expanding and ultimately joining the polar coronal holes .
they were able to simulate the formation of three coronal holes using kinematic photospheric flux transport models in combination with extrapolated potential - field source - surface ( pfss ) models for the coronal field .
they found that the coronal holes could be explained by a combination of photospheric flux transport processes : turbulent diffusion , differential rotation and the interaction between neighboring active regions .
they concluded that the global distribution of photospheric flux determined where the coronal holes formed , in particular the proximity or otherwise of strong flux of the opposite polarity . in the pfss model
, coronal loops are assumed to become open if they extend beyond a certain height , hence open field regions form wherever there is insufficient photospheric flux of the opposite polarity to connect to .
karachik et al . (
2010 ) used michelson doppler imager ( mdi ) and gong full - disk magnetograms and extreme - ultraviolet imaging telescope ( eit ) euv images to investigate four cases of isolated polar coronal holes forming on the remnants of decaying active regions .
in contrast to wang et al .
( 2010 ) , karachik et al .
( 2010 ) focused on the evolution of individual bipolar active regions .
they measured several parameters of the active regions over successive rotations including the imbalance and the compactness of the magnetic flux .
they found that less compact field tends to decay faster than more compact field , causing a gradual increase in flux imbalance in the region and the formation of a coronal hole at the location of the more compact flux .
karachik et al.s four cases occurred during the ascent of cycle 24 in 2009 and their conclusions are consistent with those of sheeley and harvey ( 1979 ) .
using extrapolated pfss models karachik et al . found that in three cases , all of them holes of positive polarity , some field rooted at the coronal hole locations connected back to the photosphere close to the north polar coronal hole .
these fields therefore appeared to play a role in the solar cycle process , evolving from toroidal active region field into poloidal field . in this paper
we follow karachik et al .
( 2010 ) by focusing on the relationship between the decaying fields of active regions and the formation of coronal holes .
we investigate which factors can determine the formation of the coronal holes including active region flux imbalance , field strength , compactness , interactions with other fields , and the structure of the coronal field represented by pfss models .
we also try to clarify the role of these fields in the solar activity cycle .
most of our examples come from the ascending phase of cycle 24 with a minority of cases representing the declining phase of cycle 23 .
the paper is organized as follows .
we describe the observational data in section [ sect : data ] .
the 14 pairs of active regions and coronal holes will be introduced and their properties summarized in section [ sect : arch ] .
pfss models will be used to aid interpretation of coronal hole formation in section [ sect : pfss ] .
we will conclude in section [ sect : conclusion ] .
full - disk images of the relative doppler shift of the ni i line at 676.8 nm are available from each of the six gong telescopes at a cadence of 1 minute , weather permitting .
gong s six stations together provide round - the - clock coverage with approximately an 87% duty cycle .
the spatial sampling of the gong images is @xmath2 pixel@xmath3 and the instrumental sensitivity is about 3 g pixel@xmath3 .
since the gong instrumentation was upgraded in 2006 , resulting in a great reduction of the zero - point error in the field measurements , synoptic maps for the radial component of the photospheric field have been constructed for every carrington rotation .
the stereo sun earth connection coronal and heliospheric investigation ( secchi ) instrument package ( howard et al . 2008 ) has four instruments : an extreme ultraviolet imager , two white - light coronagraphs and a heliospheric imager .
the extreme ultraviolet imager , euvi ( wuelser et al . 2004 ) observes in four spectral channels that span the 0.1 to 20 mk temperature range . in this paper
we use synoptic maps based on 19.5 nm images , corresponding to about 2.5 mk .
coronal holes tend to appear clearest in these maps .
we searched for decaying active regions and associated coronal holes by visually inspecting photospheric synoptic magnetograms from gong and the corresponding synoptic euv maps from the stereo a spacecraft s euvi instrument .
the gong synoptic map series begins with carrington rotation ( cr ) 2047 ( august 2006 ) whereas the stereo maps begin at cr2051 ( early 2007 ) . from pairs of gong and
stereo synoptic maps we identified 14 cases where the decay of an active region could confidently be associated with the formation of a coronal hole . these include four examples from the declining phase of cycle 23 and ten from the ascent of cycle 24 .
these are listed in table [ arlist ] and examples are shown in figures [ fig : ex3 ] and [ fig : ex6 ] .
we only consider examples where active regions could be identified over at least two consecutive rotations and where the remnant flux of the decaying active region could be associated with a newly formed coronal hole in the corresponding stereo map .
gong is a ground - based network of telescopes whereas the stereo a and b spacecraft orbits lead and lag the earth in its orbit around the sun . since their launch in august 2006 , the twin stereo spacecraft have been separating at a rate of @xmath4 per year .
the observations analyzed in this paper date from early 2007 until late 2010 , over which time the separation between stereo a and earth grew from a fraction of a degree to about 85@xmath5 , representing a time delay between a gong observation of a given longitude and the corresponding stereo a observation of that same longitude of up to 6.5 days . because of this time delay the gong and stereo maps can not be treated as being nearly simultaneous , but since we are studying the slow decay of active regions and formation of coronal holes over multiple rotations the pairs of gong and stereo synoptic maps are still applicable .
our approach differs from that of karachik et al .
( 2010 ) in that we do not estimate coronal hole magnetic fluxes by deriving coronal boundaries from stereo euv data and mapping them onto gong magnetograms . instead we study the magnetic fluxes of the decaying active regions .
[ arlist ] + .properties of active regions [ cols="<,^ , < , < , < , < , < , < , < " , ] @xmath6was the leading ( l ) or trailing ( t ) flux more dense / compact ? + @xmath7was the region initially located on the pole - ward ( p ) or equator - ward ( e ) side of the pfss streamer belt , or contained within it ( s ) ? note that none of the cycle 23 examples was initially located pole - ward of the streamer belt , and only one cycle 24 region was initially located equator - ward of the streamer belt .
+ @xmath8did the coronal hole polarity match the polar field in its own hemisphere ?
note that the coronal hole polarity matched the polar field in its own hemisphere in every case except those regions initially located equator - ward of the streamer belt .
thus there was no mismatch between a coronal hole magnetic polarity and the polar field polarity on the original active - region s side of the streamer belt .
active regions are understood to emerge from the solar interior in the form of closed loops with balanced magnetic flux but with asymmetries between the positive and negative magnetic flux distributions .
an active region is usually composed of a magnetic bipole where one polarity clearly trails the other in longitude . it is well known that active regions tend to follow a simple pattern whereby the leading magnetic fluxes of the bipoles in the north and south hemispheres have opposite polarities and that these polarities switch during each activity minimum ( hale and nicholson 1925 ) , known as hale s polarity rule . for example , the leading flux in the north / south hemisphere generally had positive / negative polarity during cycle 23 and negative / positive polarity during cycle 24 .
the active regions also tend to be asymmetric in their flux distribution : the leading flux is usually nearer the equator ( hale et al .
1919 , a rule known as joy s law ) , and is also more dense and compact than the trailing flux .
the active regions asymmetry has consequences for their decay patterns .
the trailing flux tends to decay more quickly than the leading flux and be transported pole - ward , often leaving the leading flux isolated .
the leading flux may therefore be expected to open up and form a low - latitude coronal hole under some circumstances ( harvey and sheeley 1979 ) .
the asymmetric behavior of the active regions links active and polar fields together in a single magnetic activity cycle ( babcock 1961 ) . during the ascent of an activity cycle the dominant leading polarity of a hemisphere
matches the polarity of that hemisphere s polar field , whereas during the declining phase the opposite is true .
the fact that the gong and stereo data sets include the declining phase of cycle 23 and the ascent of cycle 24 allows us to investigate whether this difference has consequences for low - latitude coronal hole formation .
table [ arlist ] summarizes the quantitative details of the 14 examples of interest in this paper .
each example spans 2 - 4 rotations .
every active region in the sample followed hale s polarity rule when first observed and all but one followed joy s law for active region tilts .
table [ arlist ] also lists the magnetic fluxes and flux densities of the regions during each rotation .
we estimate the magnetic flux of a region as follows .
we find level curves of field strength 5 g and identify by eye the curves that enclose the active region of interest . in figures [ fig : ex3 ] and [ fig : ex6 ] ( top panels ) the selected level curves for examples 3 and 6 are over - plotted in black on the gong synoptic magnetograms .
we then sum by magnetic polarity the flux contained in these curves .
also shown are the average location , flux density and compactness of each region .
we follow karachik et al .
( 2010 ) in calculating the compactness of the active region magnetic field distribution using their expression for the effective radius , @xmath9 , where @xmath10|b_r ( x , y)| / \sum |b_r ( x , y)| , \label{eq : reff}\ ] ] where @xmath11 and @xmath12 are the average pixel coordinates of the field distribution .
we calculate @xmath9 for the entire flux distribution of each region and also for the positive and negative fluxes separately .
table [ arlist ] shows that the total magnetic fluxes of all 14 regions decreased over time , consistent with decay .
the average flux densities @xmath13 generally decreased as the regions decayed but sometimes @xmath13 increased when some weak flux decayed to values below the threshold used in the analysis and was no longer included in the calculation .
the effective radii @xmath9 tended to increase in time as the flux decayed and became more diffuse , until one of the polarities decayed away .
the active regions initially had magnetic fluxes ranging from about @xmath14 mx for a late cycle 23 region to @xmath15 mx for an early cycle 24 region .
the late cycle 23 regions were larger than the early cycle 24 regions on average .
the first two regions from cycle 24 were small but the regions subsequently increased in size .
the last two regions of the 14 were the second and third largest of the study .
the larger regions tended to last longer than the smaller regions . a notable exceptions to this rule was example 5 , which was one of the smallest of the 14 but lasted four rotations .
figures [ fig : ex3 ] and [ fig : ex6 ] show gong and stereo magnetograms for two typical examples from table [ arlist ] , examples 3 and 6 .
the black contour curves over - plotted on the gong maps indicate the decaying active region fields , and the associated coronal holes are indicated by dashed white contour curves on the stereo synoptic maps .
the leading fluxes , negative in both cases , were initially denser than the trailing fluxes and the negative fluxes decayed more slowly , creating local negative flux imbalances in both cases , from which coronal holes developed as shown in the plots .
this behavior is much as expected from the previous work cited in the first paragraph of this section .
karachik et al .
( 2010 ) studied example 3 and reported the same behavior .
wang et al . (
2010 ) analyzed and simulated example 3 and found that the small leading - polarity hole that formed near the equator was balanced by the formation of a trailing - polarity hole near longitude 180 , which subsequently developed into an extension of the south polar hole ( see the arrowed features in figure 2 of wang et al . ) .
this positive - polarity hole was located in the trailing flux of the active region just to the east of the active region that contained the leading - polarity hole . in this case
the two bipoles were close enough together that their facing polarities connected to each other , leaving their outer polarities to open up .
coronal holes form under the constraint that new open flux of one polarity must always be balanced by opening up an equal amount of flux of the opposite polarity , or by closing down an equal amount of flux of the same polarity .
this process can be seen in figure [ fig : ex3 ] . of the three types of coronal hole identified by harvey and recely ( 2002 ) ,
all holes studied here appear to have belonged to the category of holes forming gradually in association with decaying active regions .
some of these regions produced minor flaring activity but , browsing through the stereo / euvi sky images , we did not find evidence of a significant eruption associated with sudden formation or development of the holes .
the decaying regions appeared to evolve steadily and gradually . in some cases
the fully - formed coronal hole would rotate onto the disk where no evidence of coronal hole formation was visible during the previous rotation .
in other cases the coronal hole evolved gradually over multiple rotations .
table [ chlist ] summarizes the qualitative behavior of all 14 regions . from this table
some clear patterns emerge . in all of the active regions except one ,
the leading polarity was stronger than the trailing polarity when the region first appeared in a synoptic map .
the lone exception , example 5 , had marginally more compact leading flux than trailing flux , and it was the leading polarity that was associated with the hole in this case .
example 5 was analyzed and simulated by wang et al .
( 2010 ) , who concluded from their models that the formation of the leading - polarity coronal hole was accompanied by some of the negative - polarity flux from the nearby polar hole closing down by connecting to the trailing flux of the mid - latitude active region via interchange reconnection , as is common during the ascending phase .
the cause of the fast decay of positive flux between the times that the cr2076 and cr2077 measurements were taken ( see table [ arlist ] ) may be related to the distribution of the fluxes according to the gong synoptic maps .
the negative flux was slightly larger but significantly less dense during cr2076 , and was composed of two concentrations to the west and south of the positive flux concentration .
there were therefore two neutral lines separating the positive and negative fluxes , which may have allowed much flux cancellation to take place during this time interval .
the positive flux may have survived the negative flux simply because there was slightly more of it to begin with .
the fact that the net flux of the region stays within the range @xmath16 mx throughout four rotations of decay suggests that flux cancellation was the dominant process in this case . in all but two cases the leading polarity had the more compact distribution . in the exceptional cases , examples 2 and 7 , the leading polarity was the more dense on average but , whereas the leading polarity decayed more slowly and opened in example 2 , it was the trailing polarity that survived and opened in example 7 .
example 7 was also analyzed and simulated by wang et al .
( 2010 ) , who explained the formation and rapid decay of this hole in terms of interchange reconnection between the trailing polarity of the active region and the polar hole boundary .
this trailing - polarity ( positive ) hole ( which can be seen in the cr2085 euvi / a map in figure 6 of wang et al . 2010 , and which we will model in section [ sect : pfss ] ) is associated with the leading negative flux of the neighboring region just eastward that produced a leading - polarity hole .
wang et al . ( 2010 ) showed the roles of diffusion and differential rotation in the formation and evolution of the two holes .
we can not generally find observations that catch the active regions at their moment of emergence when their flux is expected to be balanced .
as soon as a region emerges it interacts with its surroundings and some reconnection will inevitably take place .
according to the estimates in table [ arlist ] the earliest observed fluxes were balanced to within 10% in all regions except one .
the exception was the first example which was already a large , mature region at the beginning of the sequence of observations .
in most but not all cases the two polarities decayed asymmetrically so that one polarity dominated by the time the coronal hole appeared in the stereo data . in the two cases where the fluxes remained approximately balanced , examples 10 and 12 , the two polarities had become very distant from each other over time , resulting in an isolated unipolar region of field where a coronal hole could form . in 12 out of 14 cases the leading flux opened , whereas the denser polarity opened in only ten out of 14 cases and the sign of the initial flux imbalance matched the final flux imbalance in only eight out of 14 cases .
this indicates a strong hemispheric polarity preference in the decay rates , a pattern that is better correlated with the polarities of the coronal holes than the intrinsic asymmetries of the active regions . on the other hand ,
the polarity of the coronal hole matched the polarity of the polar field in its hemisphere in only ten out of 14 cases , showing that the hemispheric preference is not strict .
from the ten cycle 24 active regions the leading flux formed the coronal hole in eight cases .
the exceptions are examples 7 and 12
. in example 12 a coronal hole appeared in the stereo cr2091 map near the trailing flux .
this may seem surprising because in this example the leading flux was initially stronger and more compact .
this trailing flux seemed to open following the emergence of two active regions to the east of the decaying region .
the emergence of these regions may therefore have played a key role in the formation of the coronal hole from the trailing flux ( see section [ sect : pfss ] ) . in example 7
the leading negative flux was initially denser and more compact than the trailing flux , and yet the leading flux decayed faster creating a positive flux imbalance and a coronal hole
. again the photospheric context of the region will be necessary to explain the formation of this hole ( wang et al .
we will return to both examples 7 and 12 in section [ sect : pfss ] . in each of the remaining eight cases from cycle 24
the leading polarity formed the coronal hole , even when the leading flux was not the denser or more compact polarity .
the data suggest that the global field structure may have a powerful influence over the polarity distribution of coronal holes .
tables [ arlist ] and [ chlist ] show that the coronal holes tend to divide by polarity between the two hemispheres , with negative in the north and positive in the south , matching the polarities of the polar fields .
this occurred in most but not all cases : 11 out of 14 . during the ascent of an activity cycle the compact and dense polarity of a typical active region following hale s polarity rule and joy
s law would match the polarity of the polar field in its hemisphere , and the resulting low - latitude holes would therefore match the polar hole in polarity .
tables [ arlist ] and [ chlist ] hint at a link between the global field structure and the polarity of the field that forms the coronal hole . on the other hand , during the declining phase of cycle 23 this hemispheric bias seems not to have been in strong effect .
the leading fluxes of the first four examples were of opposite polarity to the polar fields in their hemispheres .
also these four active regions were relatively large and were situated closer to the equator .
one might therefore expect the relationship of these declining - phase coronal holes to the global field to be qualitatively different from their ascending - phase counterparts .
examples 1 and 3 appear to be analogous to most of the cycle 24 examples in that their leading dense fluxes decayed more slowly , creating an imbalance , and coronal holes form in association with them .
examples 2 and 4 appear to have behaved differently .
in example 2 the trailing negative flux was initially the more compact and a negative flux imbalance had developed by the next rotation .
yet , the coronal hole that formed was associated with the positive flux , the only case among the 14 whose flux imbalance and coronal hole polarities did not agree .
example 4 also showed unexpected behavior . here
the leading negative flux was initially the denser and more compact and yet it decayed faster , leaving a positive flux imbalance and a coronal hole .
these two examples seem difficult to interpret with magnetograms and euv images alone . in the following section we will revisit them using pfss models . while we only have a small number of statistics , it is apparent that there was a larger hemispheric bias in the polarities of the low - latitude holes during the ascent of cycle 24 than during the decline of cycle 23 .
the relationship between the active regions and the global coronal field structure is fundamentally different during early phases of the solar cycle compared to late phases . during early / late phases an active region typically has relatively high / low latitude and its lagging flux usually opposes / matches the polarity of the photospheric polar field in its hemisphere .
we know that during relatively quiet phases of the solar activity cycle it is the low - order multipoles , in particular the dipole and octupole , that dominate the global coronal field structure , and that these components are well correlated with the photospheric polar fields ( petrie 2013 ) . to search for possible links between the the global coronal field structure and the creation of coronal holes , information for the coronal field is needed . to this end , we will present extrapolated potential field models in the following section .
the euv images do not tell us the polarity of a coronal hole .
furthermore , as we saw in section [ sect : arch ] , a simple relationship between an euv dark region and the unbalanced magnetic flux can not be assumed to determine the polarity of a coronal hole . to supply this and other important information on the coronal magnetic field we extrapolated pfss models from each gong synoptic magnetogram . for active regions near the beginnings or ends of carrington rotations we used composite synoptic maps formed from half of the rotation containing the region and half of the previous or subsequent rotation .
extrapolating coronal pfss field models ( altschuler and newkirk 1969 , schatten et al .
1969 ) from the gong synoptic maps is a simple and effective way to diagnose the response of the coronal magnetic field to the photospheric activity patterns described in section [ sect : arch ] .
low in the corona , the magnetic field is sufficiently dominant over the plasma forces that a force - free field approximation is generally applicable .
moreover , for large - scale coronal structure the effects of force - free electric currents , which are inversely proportional to length scale , may be neglected .
we identify the lower boundary of the model with the solar photosphere at @xmath17 where @xmath18 is the solar radius .
we use the photospheric radial field maps from gong to supply the lower boundary data for the radial field component .
this use of radial field data , derived from longitudinal measurements of non - force - free photospheric fields , assuming the magnetic vector to be approximately radial in general , has been found to produce models that more successfully reconstruct coronal magnetic structure than the direct application of the longitudinal measurements as boundary data ( wang and sheeley 1992 ) . a synoptic map construction method that derives the radial component from longitudinal measurements without assuming the fields to be approximately radial has been developed and applied to chromospheric data by jin et al .
( 2013 ) with promising results .
here we adopt the standard photospheric synoptic maps because measured photospheric fields are found to be approximately radial in general ( svalgaard et al .
1978 , petrie and patrikeeva 2009 , gosain and pevtsov 2012 ) , and the resulting models remain useful and competitive for reasons explained by wang and sheeley ( 1992 ) . above some height in the corona , usually estimated to be between 1.5 and 3.5 solar radii , the magnetic field is dominated by the thermal pressure and inertial force of the expanding solar wind . to mimic the effects of the solar wind expansion on the field
, we introduce an upper boundary at @xmath19 , and force the field to be radial on this boundary by setting the potential to zero there , following altschuler and newkirk ( 1969 ) , schatten et al .
( 1969 ) and many subsequent studies .
all fields reaching @xmath20 are interpreted as being open to the heliosphere and as representing coronal holes .
the usual value for @xmath21 is @xmath22 although different choices of @xmath21 lead to more successful reconstructions of coronal structure during different phases of the solar cycle ( lee et al .
2011 ) . in this paper
we adopt the value @xmath23 because this value leads to a reasonable visual resemblance between open fields in the models and coronal hole distributions as seen in the stereo euv synoptic maps . with these two boundary conditions ,
the potential field model can be fully determined in the domain @xmath24 .
we emphasize that each pfss model is therefore completely determined by the corresponding photospheric magnetogram .
there is no time - dependence in the pfss model .
the model is successful because , to a good approximation , the global coronal field relaxes completely to a near - current - free state independent of the previous magnetic field in the sequence .
we computed the pfss models using the national center for atmospheric research s ( ncar ) mudpack package .
although pfss models can be calculated analytically , and have been so treated for several decades , we adopt a finite - difference approach in this paper to avoid some problems associated with the usual approach based on truncated spherical harmonic series ( tth et al .
mudpack ( adams 1989 ) is a package for efficiently solving linear elliptic partial differential equations ( pdes ) , both separable and non - separable , using multigrid iteration with subgrid refinement procedures .
jiang and feng ( 2012 ) have recently presented a high - speed combined spectral / finite - difference pfss solution method using the related ncar fishpack package for separable pdes . in this paper
we use the mudpack package to solve laplace s equation in its non - separable form in spherical coordinates , subject to the boundary conditions described above .
a successful pfss model tells us not only the polarity of a coronal hole but also the context of the hole within the global coronal field structure , such as its relationship to the polar coronal holes and the streamer belt .
figures [ fig : ex3 ] and [ fig : ex6 ] ( third rows of plots ) show the pfss models for examples 3 and 6 . here
selected fields and features of the model are over - plotted on the gong synoptic magnetograms .
the green and red contours indicate foot - point locations of positive and negative open fields , corresponding to coronal holes .
the blue lines delineate the tallest closed field trajectories in the solution , corresponding to the streamer belt , and the black line shows the set of apex locations of these fields , representing the equatorial current sheet
. we will refer to this black line as the source - surface neutral line ( ssnl ) in the following .
the yellow lines represent the active region fields whose foot - points lie within the black contour curves of figures [ fig : ex3 ] and [ fig : ex6 ] ( top plots ) .
the figure also shows plots of the models in spherical coordinates .
these plots show magnetic lines selected according to the field intensity at their foot - points .
closed lines are plotted in blue and positive / negative open lines in green / red .
the coordinate system is rotated so that the decaying region is at central meridian .
the plots show the development of open field at the location of the decaying region and enhanced connectivity to the southern polar latitudes .
the ssnl is generally confined to low latitudes around solar minimum , but major active regions can cause large excursions of the ssnl from the equator . during the decline of cycle 23 and the ascent of cycle 24 , the axisymmetric dipole and octupole components of the coronal field
had the strongest influence over the global coronal field structure , and these components are well correlated with the slowly - evolving photospheric polar fields ( petrie 2013 ) . during this time
, the streamer belt separated the corona into two distinct open - field regions : the global corona had negative / positive open field north / south of the ssnl .
it is striking that none of the cycle 23 coronal holes formed pole - ward of the ssnl and only one of the cycle 24 holes , example 7 , formed equator - ward , as table [ chlist ] shows .
figures [ fig : ex3 ] and [ fig : ex6 ] are illustrative of this pattern , showing the equator - ward example 3 from late in cycle 23 and the pole - ward example 6 from early in cycle 24 .
this is perhaps not surprising behavior for the declining / ascending phase of a solar cycle , when active regions emerge at relatively low / high latitudes .
the formation of a coronal hole did not result in a significant change in the location of the ssnl in any of the examples .
the low - order magnetic multipole fields , corresponding to large spatial scales , were not significantly perturbed by the photospheric active - region decay associated with the formation of the coronal holes . in each case
the hole appeared with the same polarity as the pre - existing open field on its side of the streamer belt .
this includes the exceptional case , example 7 , where the ssnl looped around the region in the northern hemisphere , placing the region in the southern half of the corona in magnetic terms . in a few late cases , examples 12 - 14 , when the sun was slightly more active and the regions were more numerous and slightly larger , the corona had a more complex structure and the ssnl overlay the active regions . under these conditions
the decay rates of the positive and negative polarities seem to have been determined by their relative density and compactness .
examples 1 - 4 occurred during the declining phase of cycle 23 and were all located near the equator .
two of these holes ( examples 1 and 3 ) appeared equator - ward of the ssnl and two of them ( examples 2 and 4 ) were straddled by it .
that none of them occurred pole - ward of the ssnl signifies an important difference between the cycle 23 and cycle 24 coronal holes .
table [ chlist ] shows that of the ten out of 14 cases where the polarity of the coronal hole matched the sign of the polar field in its hemisphere , six out of six matched for pole - ward cases and zero out of three for equator - ward cases .
these statistics point to a strong influence of the global field structure on the polarity of successfully formed coronal holes . during cycle 23
the leading flux did not match the polarity of the polar field in its own hemisphere , making it much more difficult for it to open and form coronal holes pole - ward of the ssnl than equator - ward . of the equator - ward cases , in the cycle 23 examples 1 and 3 , the leading flux had the same polarity as the polar field on the opposite hemisphere ,
i.e. , the same polarity as the polar field on its side of the streamer belt .
these holes were therefore topologically equivalent to cycle 24 holes pole - ward of the streamer belt .
a visual comparison of example 3 in figure [ fig : ex3 ] and example 6 in figure [ fig : ex6 ] illustrates this topological relationship .
the major difference between these two examples is that example 3 occurred south of the equator . in terms of magnetic topology
they are equivalent . in common with the cycle 24 holes , the equator - ward cycle 23 holes in examples 1 and 3 formed from flux of the same polarity as the pre - existing open field on their side of the streamer belt . in these two cases
the leading flux was also more dense and more compact than the trailing flux .
the global field structure and the magnetic asymmetry of the active region together seem to have enabled the leading - polarity coronal hole to form . of the five cases occurring near the ssnl ,
the denser polarity formed the coronal hole in only three of them : zero out of two during cycle 23 and three out of three during cycle 24 .
the three cycle 24 cases appear to have evolved straightforwardly according to their asymmetric flux density distributions .
the denser leading polarity survived to form the hole in all three cases . as noted in the previous section , the two examples from cycle 23 , examples 2 and 4 , are more difficult to explain .
aided by the pfss models , we do so now .
example 2 was centered a degree or two north of the equator and had the leading positive polarity of a northern - hemisphere cycle 23 region .
the ssnl straddled the active region and the trailing negative flux was initially more compact than the positive flux . a significant negative flux imbalance developed as the positive flux decayed more quickly than the negative flux .
according to the pfss models this flux imbalance manifested mostly as increased magnetic connection to the positive south polar latitudes .
the coronal hole that did form near the region had positive polarity even though the flux imbalance of the decaying region was clearly negative ( -34% ) .
a newly emerged region to the west visible in the cr2053 map ( not shown here ) seems to have been involved in the formation of this coronal hole .
its positive trailing polarity combined with the leading polarity of the decaying region to create an area of positive flux where a coronal hole could develop .
this is the only example in our sample of 14 where the polarities of the decaying active region flux imbalance and the associated coronal hole did not match . as in example 2 , in example 4 ( figure [ fig : ex4mod ] ) the ssnl straddled the active region but this southern region had leading negative flux that was initially much denser and more compact than its positive flux . and yet the flux balance became overwhelmingly positive over time and a coronal hole of positive polarity formed .
in this case the positive flux imbalance and associated asymmetric topology seem to explain the asymmetric development of this region .
the active region was initially tucked underneath the streamer belt .
the streamer belt separated the closed flux underneath from the negative open flux to the north and the positive open flux to the south . based on its positive flux imbalance , this region
might be expected to have had more magnetic connection to the north ( negative ) side half of the corona than the south , consistent with its positive flux imbalance . according to the model
, most of these connections occurred within the streamer belt via lines arching from the positive flux of the region over the western half of the region to the distant negative flux of the neighboring region to the west .
these connections were discussed in detail by petrie et al .
( 2009 ) and are evident in figure [ fig : ex4mod ] . since the connecting fields extended high in the streamer belt they were particularly susceptible to being reconnected at the ssnl and opened up . according to the plots for the subsequent rotations shown in figure [ fig : ex4mod ] , this seems to be what happened . by the time of cr2070 ,
some of the positive flux of the region was open and a dark coronal hole corresponding to the positive flux was apparent in the euv map , but the basic global structure of the coronal field was unchanged .
figure [ fig : ex5mod ] shows the stereo euv maps and pfss models for example 5 .
as noted in the last section , this small region survived several rotations and underwent fast decay between crs 2076 and 2077 , perhaps due to flux cancellation . a sizable coronal hole developed to the west of its main leading - polarity ( negative ) flux concentration between crs 2076 and 2077 .
this coronal hole stands out because it developed in association with negative active region flux that was less dense than the corresponding positive flux in the region and yet decayed more slowly and survived to create the coronal hole .
we discussed in the last section that this may have been due to the distribution of the negative flux in two concentrations , bordering the positive flux at two neutral lines , visible in figure [ fig : ex5mod ] allowing flux cancellation to dominate the evolution during crs 2076 and 2077 .
given that there was slightly more negative than positive flux observed in this region from the start of the sequence of gong observations , flux cancellation alone would ultimately produce a significant negative flux imbalance .
the pfss models show that the resulting negative coronal hole also preserved the global coronal structure as a positive hole in this region could not have done .
we found no example where the formation of a coronal hole significantly changed the coronal structure .
the pfss model plots for example 7 in figure [ fig : ex7mod ] demonstrate the influence active regions can have on the global structure of the corona .
the ssnl followed a tight loop - shaped trajectory to straddle the active region whose trailing positive polarity ultimately produced the coronal hole . throughout the three rotations the ssnl maintained this shape and separated this open trailing positive flux from the open leading negative flux of the neighboring region to the east .
the ssnl appears to have been locked in this position as the active regions decayed and the euv coronal hole formed from this trailing flux .
this is in marked contrast to the effect of this region when it emerged into the corona .
as figure [ fig : ex7emer ] shows , there was a major change between cr2082 , when the ssnl lay south of the equator around @xmath25 longitude before the region emerged , and cr2083 , when the ssnl encroached more than @xmath26 northward into the northern hemisphere in response to the emergence of the region and its neighbor to the east .
the contrasting responses of the global field structure to active region emergence and decay are due to the effects of these processes on the photospheric field distribution and their timescales . when active regions emerge into the atmosphere at a location where no pre - existing active - region flux is present .
the effect on the global field can be significant .
the low - order multipole moments of the active region can be large enough to perturb the streamer belt over timescales much less than a rotation as in figure [ fig : ex7emer ] .
the effects of the decay of this region was much slower : between crs 2083 and 2085 the region decayed almost completely yet the streamer belt structure hardly changed .
the decay processes left a large - scale patch of weak positive flux which , with the large - scale patch of weak negative flux left by the decay of the neighboring region to the east , maintained a low - order magnetic multipole moment in cr 2085 comparable to that of the newly emerged region in cr 2083 .
this preserved the kink in the streamer belt even after the active - region - strength fields had almost completely decayed away . in this case
the streamer belt did not lose its kink until cr2087 , by which time the weak unipolar photospheric fluxes of the two regions had been transported from the area . in each of our examples the remnant flux of the decayed active region
generally maintained the influence of the region on the global coronal field at least a rotation after the region had decayed .
in several of the examples , in particular 1 , 2 , 7 and 12 , we see evidence in the models that the active region field was open long before a dark coronal hole was observed in euv images , where the coronal hole did not become visible until the active region decayed .
the plots of the models for example 7 , shown in figure [ fig : ex7mod ] , and example 12 indicate that the active regions can contain flux that is open but does not appear as dark regions in the euv images .
luhmann et al .
( 2009 ) showed evidence of contributions to the solar wind from low - latitude coronal holes during early 2007 , including the two active regions from examples 1 and 2 .
in example 12 , according to the models , the leading flux was partially open during both crs 2090 - 1 but did not appear in the euv observations as a dark region .
meanwhile a dark coronal hole appeared in euv in association with the decaying trailing flux as mentioned in the previous section .
the overall behavior of the decaying active regions may be summarized as follows .
the structure of the global coronal field , including the streamer belt , was preserved during the decay of the active regions and the formation of the coronal holes .
thus where the active region was decisively on one side of the streamer belt or the other , the dominant polarity of the open field on that side of the streamer belt invariably matched the polarity of the active - region field that formed the hole . in all cases of isolated active regions within the steamer belt , straddled by the ssnl ,
the intrinsic magnetic asymmetry of the region became important : the denser , more compact polarity was the one that formed the coronal hole . in cases where the region was within the streamer belt and
was not magnetically isolated the coronal hole formed in a complex way particular to the magnetic context of the region , but always without changing the global magnetic structure of the corona .
a typical late - phase active region is located closer to the equator and is larger than an ascending phase region , making the structure of the global field somewhat less influential in its topological development .
the intrinsic properties of the active region and its complex interactions with neighboring regions therefore appear to be more important factors in determining the behavior of declining - phase regions than of ascending phase regions .
nevertheless the global field structure did not change significantly as a result of the decay of even the largest declining - phase regions studied here .
only the emergence of new active regions seem to have produced significant global change .
active region emergence can significantly change the global structure of the corona , reshaping the streamer belt , and sometimes creating multiple streamer structures . even under the relatively quiet conditions prevailing between 2007 and 2010
when the low - order multipoles dominated the global coronal field , major new active regions produced major changes in the global coronal structure ( see figure [ fig : ex7emer ] ) .
yet we have found that active region decay produces no significant change in the the global structure of the corona over multiple rotations , even for large active regions that develop major flux imbalances .
the active photospheric fields change significantly from rotation to rotation , as the figures in this paper show .
a carrington rotation is plenty of time for the coronal field to adjust to a changed photospheric field .
furthermore the pfss model for a given rotation is computed completely independently of the previous rotation .
it is therefore striking that the pfss coronal field models show almost global structure change due to coronal hole formation over multiple rotations .
two facts seem to explain the continuity of global coronal field structure during coronal hole formation . first , the global coronal field varied gradually over the period of time studied . during the years 2007 - 2010 the low - order multipoles ,
in particular the dipole and octupole , had the strongest influence over the global coronal field structure , and these components are well correlated with the photospheric polar fields ( petrie 2013 ) .
the stability of the large - scale component of the global coronal field is due mostly to their relationship to the slowly - evolving polar fields .
the resulting large - scale dipolar and octupolar coronal fields were generally more globally influential than the higher - order coronal fields associated with decaying active regions .
photospheric flux transport processes produce unipolar bodies of flux of both polarities in both hemispheres , but the coronal fields of decaying active - regions did not have large enough low - order multipole moments or global influence to cancel the dominant multipoles .
they could therefore only open under the condition that they did not significantly change the basic coronal structure dominated by these lower - order fields .
unipolar fields of the `` wrong '' polarity generally remained closed ( and unstudied ) .
thus the basic global coronal field structure , dominated by low - order , large - scale , slowly - varying fields , was preserved .
second , the remnant weak field had a long - lasting effect that remained after each region had decayed away .
decaying regions that produced coronal holes tended to leave large - scale patches of unipolar weak field which took several rotations to disperse after the region itself had decayed away .
these remnant fields were much weaker than the original active - region fields but they were spread over a much larger area .
the weak remnant fields had low - order multipole moments of comparable strength to those of the original active regions , and thereby approximately preserved the global field signature of the active region long after the region itself has decayed away .
this is why the global coronal field structure remained almost unchanged during the process of coronal hole formation .
we can also comment on the role of these decaying regions in the solar activity cycle .
in babcock - leighton models for the solar cycle ( see , e.g. , charbonneau 2010 ) , the global solar field is often idealized as an approximately axisymmetric configuration , with a belt of tilted active regions in each hemisphere and a global coronal ssnl that does not travel far from the equator . in such models
the more dense , compact leading active - region fluxes in the two hemispheres interact with each other across the equator while their less dense trailing counterparts decay and are transported pole - ward .
both the leading and trailing fluxes would contribute to the activity cycle in converting active - region toroidal flux into poloidal flux : the leading fluxes would form trans - equatorial poloidal loops across the equator and the trailing flux would join the nearly axisymmetric polar flux .
it is perhaps not surprising that the fields studied in this paper do not behave in this simple way .
the photospheric field rarely featured matching pairs of active regions in the two hemispheres .
when the global corona had a simple ssnl and a nearly axisymmetric appearance the active regions were small and distant from the equator , and they did not interact significantly with fields on the opposite side of the streamer belt .
their unbalanced flux simply opened up , converting nearly - toroidal active - region flux to nearly - poloidal open flux . in large , magnetically isolated region within the streamer belt the less dense and compact polarity of the region tended to decay faster than the other polarity , leaving it unbalanced and with no partner in the opposite hemisphere to connect with .
when a major flux imbalance developed in a region some of its flux necessarily connected elsewhere , usually converting nearly - toroidal active region flux into the nearly - poloidal field of the streamer belt or the open fields .
in these ways the decaying active regions participated in the solar activity cycle .
we thank the referee and alexei pevtsov for helpful comments .
k.j.h . carried out this work through the national solar observatory research experiences for undergraduate ( reu ) site program , which is co - funded by the department of defense in partnership with the national science foundation reu program .
this work utilizes data obtained by the nso integrated synoptic program ( nisp ) , managed by the national solar observatory , which is operated by aura , inc . under a cooperative agreement with the national science foundation .
the data were acquired by instruments operated by the big bear solar observatory , high altitude observatory , learmonth solar observatory , udaipur solar observatory , instituto de astrofsica de canarias , and cerro tololo interamerican observatory .
the stereo / secchi data used here are produced by an international consortium of the naval research laboratory ( usa ) , lockheed martin solar and astrophysics lab ( usa ) , nasa goddard space flight center ( usa ) rutherford appleton laboratory ( uk ) , university of birmingham ( uk ) , max - planck - institut fr sonnensystemforschung ( germany ) , centre spatiale de lige ( belgium ) , institut doptique thorique et applique ( france ) , institut dastrophysique spatiale ( france ) .
adams , j. 1989 , applied math . and comput .
34 , 113 altschuler , m. , newkirk , g. 1969 , solar phys . 9 , 131 babcock , h.w .
1961 , apj133 , 572 charbonneau , p. 2010 , living rev .
solar phys . 7 , 3 cranmer , s.r .
2009 , living reviews in solar physics 6 , 3 de toma , g. 2011 , solar phys .
274 , 195 hale , g.e . ,
nicholson , s.b .
1925 , apj 62 , 270 hale , g.e . ,
ellerman , f. , nicholson , s.b . , joy , a.h .
1919 , apj 49 , 153 hoeksema , j.t .
1984 , phd thesis , stanford univ .
howard , r.a . ,
moses , j.d .
, vourlidas , a. , newmark , j.s . ,
socker , d.g . ,
plunkett , s.p .
2008 , ssrv 136 , 67 jin , c.l . ,
harvey , j.w . & pietarila , a. 2013 , apj 765 , 79 jones , h.p .
2005 , large - scale structures and their role in solar activity ( asp conf .
346 ) , ed .
k. sankarasubramanian , m. penn , & a. pevtsov ( san francisco , ca : asp ) , 229 kaiser , m.l . ,
kucera , t.a .
, davila , j.m . , st .
cyr , o.c .
, guhathakurta , m. & christian , e. 2008 , ssrv 136 , 5 karachik , n.v . ,
pevtsov , a.a .
& abramenko , v.i .
2010 , apj 714 , 1672 lee , c.o . ,
luhmann , j.g . ,
hoeksema , j.t . ,
sun , x. , arge , c.n .
& de pater , i. 2011 , solar phys .
269 , 367 lee , c.o . ,
luhmann , j.g . ,
zhao , x.p . ,
liu , y. , riley , p. , arge , c.n . ,
russell , c.t . , de pater , i. 2009 , solar phys . 256 , 345 luhmann , j.g . ,
lee , c.o . , li , y. , arge , c.n . ,
galvin , a.b .
, simunac , k. , russell , c.t . , howard , r.a . & petrie , g. 2009 , solar phys . 256 , 285 mackay , d. & yeates , a. 2012 , living reviews in solar physics 9 , 6 parker , e.n .
1958 , apj 128 , 664 petrie , g.j.d .
2013 , apj 768 , 162 petrie , g.j.d . , canou , a. , amari , t. 2009 , solar phys .
274 , 163 petrie , g.j.d . & patrikeeva , i. 2009 , apj 699 , 871 schatten , k. , wilcox , j. , ness , n. 1969 , solar phys . 6 , 442 sheeley , n.r .
, jr . 2010 , in cranmer , s.r . , hoeksema , j.t . ,
kohl , j.l .
, soho-23 : understanding a peculiar solar minimum , asp conf . ser .
428 , 3 thompson , b.j . ,
gibson , s.e . , schroeder , p.c . ,
webb , d.f .
, arge , c.n .
, bisi , m.m . , et al .
2011 , solar phys .
274 , 29 tth , g. , van der holst , b. , huang , z. 2011 , apj 732 , 102 wang , y .- m . ,
robbrecht , e. , rouillard , a.p . ,
sheeley , n.r .
thernisien , a.f.r .
2010 , apj 715 , 39 wang , y .- m . , sheeley , n.r . jr . 1992 , apj 392 , 310 wuelser , j .- p . , lemen , j.r . ,
tarbell , t , d . ,
wolfson , c.j . ,
cannon , j.c . , carpenter , b.a .
2004 , spie 5171 , 111 | we study the relationship between decaying active region magnetic fields , coronal holes and the global coronal magnetic structure using global oscillations network group ( gong ) synoptic magnetograms , solar terrestrial relations observatory ( stereo ) extreme ultra - violet ( euv ) synoptic maps and coronal potential - field source - surface ( pfss ) models .
we analyze 14 decaying regions and associated coronal holes occurring between early 2007 and late 2010 , four from cycle 23 and 10 from cycle 24 .
we investigate the relationship between asymmetries in active regions positive and negative magnetic intensities , asymmetric magnetic decay rates , flux imbalances , global field structure and coronal hole formation . whereas new emerging active regions caused changes in the large - scale coronal field , the coronal fields of the 14 decaying active regions only opened under the condition that the global coronal structure remained almost unchanged .
this was because the dominant slowly - varying , low - order multipoles prevented opposing - polarity fields from opening and the remnant active - region flux preserved the regions low - order multipole moments long after the regions had decayed .
thus the polarity of each coronal hole necessarily matched the polar field on the side of the streamer belt where the corresponding active region decayed .
for magnetically isolated active regions initially located within the streamer belt , the more intense polarity generally survived to form the hole . for non - isolated regions ,
flux imbalance and topological asymmetry prompted the opposite to occur in some cases .
= 1 |
it is well known that the sun is an unique star which affect the geo - space environment extremely .
any variations of solar activities may greatly affect many sides of human living , from satellite operations , radio - based communication , navigation systems , electrical power grid , and oil tubes , etc .
therefore , it is very important to understand when and how the solar activity will take place in the near future .
there are many people who had made such endeavors ( hiremath , 2008 ; hathaway , 2009 ; strong , julia , & saba , 2009 ; etc ) . however , we have much of uncertainties ( pesnell , 2008 ) to present the details of the forthcoming solar cycles . the film * 2012 *
reflects public misgivings to the imminent impact of solar fierce eruptions around the year of 2012 .
there are many methods to predict the solar cycle 24 and the beyond .
the solar cycle 24 prediction panel in october 2006 made a comprehensive list of prediction methods including precursor , spectral , climatology , recent climatology , neural network , physics - based , etc ( pesnell , 2008 ) . however , the investigation of periodicity of solar activity is a groundwork for solar cycle prediction . in previous works ,
many kinds of periodic modes are found from the analysis of different solar proxies ( e.g. , yearly averaged sunspot number , cosmogenic isotopes , historic records , radiocarbon records in tree - rings , etc . ) , for example , the 11-yr solar cycle , 53-yr period ( le & wang , 2003 ) , 78-yr period ( wolf , 1862 ) , 80 - 90-yr period ( gleissberg , 1971 ) , the 65 - 130 yr quasi - periodic secular cycle ( nagovitsyn , 1997 ) , 100-yr periodic cycle ( frick et al , 1997 ) , 101-yr period cycle ( le & wang , 2003 ) , 160 270-yr double century cycle ( schove , 1979 ) , 203-yr suess cycle ( suess , 1980 ) .
otaola & zenteno ( 1983 ) proposed that long term cycles within the range of 80 100 and 170 180 yr are existed certainly .
sunspot number is the most commonly predicted index of solar activity .
the number of solar flares , coronal mass ejections , and the amount of energy released are well correlated with the sunspot number .
the present work mainly applied the annual averaged relative sunspot number ( asn ) during 1700 2009 to study the long - term variations , and found some meaningful insights of the forthcoming solar solar activities .
this paper is arranged as following : section 2 presents the investigation of periodicity of solar activity and their main features .
section 3 gives some implications of solar cycles , such as the extrapolation of the forthcoming schwabe cycle 24 , the distribution of the solar flares , and the possible origin of solar cycles .
section 4 is the conclusions and some discussions .
it has been realized for a long time that the solar activity is connected tightly with the solar magnetic field .
the energy released in solar eruptive processes is coming from the magnetic field .
hence the variations of solar magnetic field will be a physically reasonable indicator for the solar activity .
the sunspot number presents a perceptible feature of the solar magnetic field , which can be regarded as a commonly predicted solar activity index .
so , in this work , we adopt asn during 1700 2009 to investigate the periodicity of the solar long period activities .
the data set is downloaded from the internet : http://sidc.oma.be / sunspot - data/. the upper panel in fig.1 presents the profile of asn during 1700 2009 , and the lower panel is a fourier power spectra of the fast fourier transformation ( fft ) from asn . here presents 3 obvious spectral peaks which indicate that the strongest periodicity is occurred at 11 years ( @xmath1 ) , the second strongest periodicity is at 103 years ( @xmath2 ) , and the mildly periodicity is at a period of 51.5 years ( @xmath3 ) . the most remarkable feature in fig.1 is the solar cycles with about 11-yr periods , it is the well - known schwabe cycle . as asn reaches to the minimum 2.9 in 2008 , and then increases slightly to 3.1 in the year of 2009 .
after coming into 2010 , it becomes stronger obviously than the past two years , there occurred several goes m - class flares in the first two months .
so we may confirm that the end of solar cycle 23 and the start of solar cycle 24 is occurred around 2008 .
there are 28 schwabe cycles during 1700 2009 .
they are numbered by an international regulation since 1755 .
the solar cycle 1 started from the year of 1755 , the solar cycle 23 ended around 2008 . during 1700
1755 , there are 5 solar schwabe cycles .
we may numbered them as @xmath4 , @xmath5 , @xmath6 , @xmath7 , and @xmath8 respectively , which presented in the upper panel of fig.1 . in order to investigate the profile of each solar schwabe cycle more detail to get more useful insights
, we may define several parameters : \(a ) rise - time ( @xmath9 ) , defined as from the starting minimum to the maximum of a solar cycle in unit of year ; \(b ) decay - time ( @xmath10 ) , defined as from the maximum to the next minimum of a solar cycle in unit of year ; \(c ) the profile of solar schwabe cycles is always asymmetric , i.e. , taking less time to rise to the maximum than reaching to the next minimum .
we may define an asymmetric parameter as ( @xmath11 ) . when @xmath12 , the profile is symmetric ; @xmath13 is left asymmetric , @xmath14 is right asymmetric ; \(d ) maximum asn ( @xmath15 ) , which may represents the asn amplitude of each schwabe cycle .
table 1 lists all parameters of schwabe cycles since 1700 .
the last two lines in the table present the mean value and the standard deviation of above parameters .
.list of characteristic parameters of solar schwabe cycles during 1700 2009 .
@xmath16 , @xmath9 , @xmath10 , p , @xmath17 , @xmath18 is the start time , rise - time , decay - time , period , asymmetric parameter , and the maximum asn , respectively .
av and dev are the averaged value and the standard deviation of the related parameters , respectively .
the star @xmath19 marks the bottom of the each secular cycle . [ cols="<,^,^,^,^,^,^,^",options="header " , ] table 1 indicates that the rise - time of schwabe cycles is in range of 3 7 years ( the longest rise - time is 7 years in cycle 7 ) , with the averaged length of 4.4 years , and the standard deviation of 1.2 years ; the decay - time of schwabe cycles is in range of 3 11 years ( the longest decay - time is 11 years in cycle 4 ) , with the averaged length of 6.6 years , and the standard deviation of 1.2 years ; the period of schwabe cycles is in range of 9 14 years , with the averaged length of 11 years , and the standard deviation of 1.3 years .
the distribution of asn amplitudes of schwabe cycles shows clearly the existence of 103-yr cycle , we call such long - term cycle as secular cycle .
gleissberg firstly implied the existence of secular solar cycle , so we also call it as gleissberg cycle ( gleissberg , 1939 ) .
in fact , the secular cycle is consistent with the appearance rate of the grand minima , such as the sp@xmath20rer minimum around 1500 , the maunder minimum around 1700 ( 1645 1715 , eddy , 1983 ) , the dalton minimum around 1800 ( 1790 1820 ) . from the distribution of asn amplitudes of schwabe cycles
we may also find that around 1900 ( schwabe cycle 14 ) seems also a grand minimum .
in other word , each of the grand minima is possibly occurred in a vale between two secular cycles . in order to show clearly the secular cycles
, we make a fitted empirical function with sinusoidal shape by using a method similar to the square - least - method : at first we assume the sinusoidal function in following formation : @xmath21 here , @xmath22 represents the time from 1700 ( with unit of year )
. then let the following sum to become minimum : @xmath23^{2}\rightarrow min.\ ] ] in above equations , @xmath24 and @xmath25 present the values of the fitted empirical function and the observations of asn , respectively .
as equation ( 1 ) is a nonlinear function , we could nt obtain the true values of parameters a , b , c , d , and e from the standard square - least - method .
however , we may try to list a series of [ a , b , c , d , e ] and calculate the values of @xmath26 , respectively .
then find out the minimum value of @xmath26 and the corresponding parameters of [ a , b , c , d , e ] .
the thick dot - dashed curve in the upper panel of fig.1 is the fitted empirical function , which can be expressed as : @xmath27 + 0.06y.\ ] ] from the empirical function we may find that the period of long - term variation of asn is 103 years , which is very close to the secular cycle .
we may numbered the secular cycles as @xmath28 ( includes the schwabe cycle e a , and 1 5 ) , @xmath29 ( includes the schwabe cycle 6 14 ) , @xmath30 ( includes the schwabe cycle 15 24 ) , and @xmath31 ( after the schwabe cycle 24 ) since 1700 , respectively marked in fig .
1 . at present
the sun is in a vale between @xmath30 and @xmath31 .
the last term in the right hand of equ.(3 ) implies that secular cycles have a gradually enhancement , and the sun may have a tendency to become more and more active at the timescale of several hundred years .
many previous studies also presented the evidences of the secular cycles ( nagovitsyn , 1997 ; frick et al , 1997 ; le & wang , 2003 ; bonev , penev , & sello , 2004 ; hiermath , 2006 , etc . ) .
the grand minima ( e.g. , sp@xmath20rer minimum , maunder minimum , dalton minimum , etc ) implies that the sun might have experienced the dearth of activity in its evolutionary history . and
there is no complete consensus among the solar community whether such grand minima are chaotic or regular .
this work may confirm the periodicity of the solar secular cycles . from the upper panel of fig.1
we can not get the obvious evidence of the cycle with period of 51.5 yr . however , the evidence is very strong in the fourier power spectra in lower panel of fig.1 .
in fact , from the upper panel of fig .
1 we may find something of that , around each peak of the secular cycle , the schwabe cycles are in mildly weak amplitudes in asn .
secular cycles seem to segment into two sub - peaks .
for example , the solar schwabe cycle a and 1 around the peak of @xmath28 , the solar schwabe cycle 10 around the peak of @xmath29 , and the solar schwabe cycle 20 around the peak of @xmath30 .
possibly , these facts are the indicator of the existence of @xmath3 component .
le & wang ( 2003 ) investigated the wavelet transformation of asn series from 1700 2002 , and found the evidence of solar cycles with period of 11-yr , 53-yr , and 101-yr . in this work ,
the asn series is spanned from 1700 2009 , and we rectify the periods as 11-yr , 51.5-year , and 103-yr .
our results are very close to that of le & wang ( 2003 ) .
additionally , we find that there is an interesting phenomenon : @xmath32 is obviously departed from any integers ; however , @xmath33 is fitly equal to an integer of 2 .
this evidence shows that @xmath2 and @xmath3 are originally connected with each other , and @xmath3 seems to be a second harmonics of @xmath2 . however , there is no such relationships between @xmath1 and @xmath2 , @xmath3 .
table 1 presents another interesting feature : most of the asymmetric parameters are less than 1.00 ( there are 22 schwabe cycles with @xmath34 among the total 28 cycles , and the proportion is 78.5% ) , the averaged value of asymmetric parameters is about 0.722 , and this implies that most schwabe cycles are left asymmetric .
they have rapidly rising phases and slowly decay phases . and
the cycle evolution can not be modelled by some simple amplitude - modulated sinusoids .
such phenomenon is called as waldmeier effect ( 1961 ) . among the total 28 cycles ,
there are only 3 schwabe cycles with symmetric profiles , and 3 schwabe cycles with right asymmetric profiles .
the symmetric cycles ( no.d , no.5 , and no.16 ) are very close to the vale of the secular cycles .
the no.7 cycle is a bizarrerie for the super long rise - time ( 7 years ) and super short decay - time ( 3 years ) with a super large asymmetric parameter ( @xmath35 ) .
additionally , there is an anti - correlated relationship between asymmetric parameter and maximum asn among 28 solar schwabe cycles .
the correlate coefficient is -0.55 .
i.e. the stronger the solar schwabe cycle , the more left asymmetric the cycle profile .
it is very important to forecast the forthcoming solar cycles .
many people make great efforts on this problem ( pesnell , 2008 ; hiremath , 2008 ; wang et al , 2009 ; strong & saba , 2009 ; hathaway , 2009 ; etc ) . according to the main features of long - term solar activity periodicity
, we may also make an extrapolation of the forthcoming solar cycles .
firstly , we note that solar schwabe cycle e , 5 , and 14 are located in the bottom of secular cycles , and have rise - time of 46 years . at the same time , the solar schwabe cycle 24 is also possibly seated in a bottom between @xmath30 and @xmath31 .
then , it is reasonable to suppose that the rise - time of the solar schwabe cycle 24 will be in the range of 4 6 years , i.e. , its maximum will occur around 2012 2014 .
the another important fact is that all durations of the three bottom solar schwabe cycle ( cycle e , cycle 5 , and cycle 14 ) are 12 years , which are relatively long duration schwabe cycles .
based on the similarity , we may deduce that the length of schwabe cycle 24 will also last for about 12 years , which is longer a bit than the normal periods of solar schwabe cycles . according to variations of the magnitude of solar cycles ( expressed as asn )
, we can make some extrapolations for the forthcoming solar cycles . from table 1 , we know that the averaged magnitude of asn among the 28 solar schwabe cycles during 1700 2009 is 105.9 with a standard deviation of 37.8 , and the three bottom cycles ( cycle e , 5 , and 14 ) have the asn magnitude from 47.5 to 63.5 . from the above investigations we know that solar schwabe cycle 24 may also be a bottom cycle
, then we have no other reason to suppose that solar schwabe cycle 24 will have an asn magnitude of exceeding the range of 47.5 63.5 , it is a really relative weak solar schwabe cycle . however , this value is lower than the most results listed in the work of pesnell ( 2008 ) , but it is consistent with the result of badalyan , et al ( 2001 ) .
2 plotted the extrapolation of the forthcoming solar schwabe cycles by the similarity of the trend - line induced from fig.1 .
it shows that the position of solar schwabe cycle 24 is very close to the valley between the secular cycle @xmath30 and @xmath31 , which is very similar to the solar schwabe cycle 5 in the valley between @xmath28 and @xmath29 , and the solar schwabe cycle 14 in the valley between @xmath29 and @xmath30 .
so it is very reasonable to extrapolate that the solar schwabe cycle 24 will be similar to that of the solar schwabe cycle 5 and cycle 14 .
this similarity implies that the schwabe cycle 24 will be a relatively weak activities .
based on these similarities and equ.(1 ) , we may plot the extrapolated asn of the subsequent solar schwabe cycles and the secular cycle @xmath31 in fig.2 . from these extrapolations
, we may find that solar schwabe cycle 24 will reach to the apex in the year of about 2012 - 2014 , and this cycle may last for about 12 years or so .
it will be a long , but relatively weak schwabe cycle . when and what kind of solar flare events will occur is also an intriguing problem .
fig.3 presents comparisons between asn and the distributions of the appearance rate of solar flares and the annual averaged solar radio flux ( arbitrary unit ) at frequency of 2.84 ghz around the solar schwabe cycle 23 .
the appearance rate of c - class , m - class , and x - class ( limit from x1.0 to x9.9 ) flares are presented by their annul flare numbers , and the appearance rate of the super - flares ( here we define the super - flare as @xmath0 x10.0 ) are presented by their related goes soft x - ray class .
the vertical dashed line marks the time of magnetic maximum of solar schwabe cycle 23 .
the 2.84 ghz solar radio flux is observed at chinese solar broadband radiospectrometer ( sbrs / huairou ) ( fu et al , 2004 ) .
the data set of the annul flare numbers is compiled from goes satellite at soft x - rays ( http://www.lmsal.com/sxt/homepage.html ) .
from fig.3 we may find an obvious tendency : the more powerful solar flares are inclined to occur in the period after the maximum of solar schwabe cycle ( presented by the two tilted - dotted lines ) .
for example , the asn of solar schwabe cycle 23 reaches to its maximum in about 2000 , and the annual number of c - class flares and m - class flares reach to the maximum in about 2001 , while the annual number of x - class flare reaches to its apex in about 2002 , the most concentration of the super - flares is occurred in 2003 . during the whole schwabe cycle 23 ( 1996 2008 ) there are 12995 c - class flares , 1444 m - class flare , 119 x - class flares , and 6 super - flares . among them
there are 7414 c - class flares ( 57.1% ) , 950 m - class flare ( 65.8% ) , 80 x - class flares ( 67.2% ) and all the super - flares are occurred after the year of asn apex ( 2000 ) of the cycle . in 2005 , which is very close to the magnetic minimum of cycle 23 , there are 18 x - class flares occurred .
additionally , the annual averaged solar radio flux ( in panel b ) at frequency of 2.84 ghz are very similar to the profile of the annual c - class flare numbers .
generally , the solar radio emission at 2.84 ghz is mainly associated with the non - thermal eruptive processes .
it shows that all of the 6 super - flares ( @xmath36 ) occurred in the decay phase of the schwabe cycle 23 , including the the largest flare event ( x28 , 2003 - 11 - 03 ) recorded in noaa so far .
these facts imply that the stronger flare events are inclined to occur in the decay phase of the schwabe cycle .
however , so far we can not make a doubtless conclusion because we have no enough reliable observation data of the other schwabe cycles .
the above characteristics of the solar flare distribution in solar schwabe cycles is also implied the asymmetric properties of the schwabe cycles .
however , it is most intricate that most of the powerful flares are occurred not in the rising - phase but in the decay - phase . in this work
, we make an assumption that future behaviors of the solar activity in upcoming several decades years can be deduced from the averaged behaviors in the past several hundred years . and
this can be accepted because the several hundred - year is so much short related to the life - time of the sun when it is at the main sequence .
we may regard the sun as a steady - going system which will run according to its averaged behaviors of its past several hundred years and last for several hundred years again .
the time - scales of the solar 11-yr schwabe cycle and 103-yr secular cycles are much shorter than the diffusion time - scale of solar large - scale global magnetic field structures ( about @xmath37 yr , stix , 2003 ) , and much larger than the solar dynamical time - scale ( for example , the 5-min oscillations , etc ) .
the origin of these cycles is a remarkable unsolved tantalizing problem .
we need to explore some theoretical models which may carry the energy from the interior to the surface and release in solar atmosphere in some periodic forms . presently , there are two kinds of theoretical models for solar cycles , one is turbulent dynamo models , the other is mhd oscillatory models
. however , both of them have much difficulties to interpret the main features of solar cycles reliably ( hiremath , 2009 ) . as for the 11-yr schwabe cycle , in spite of much of difficulties , possibly dynamo theory is the most popular hypothesis to explain the generations ( babcock , 1961 ; leighton , 1964 , 1969 ) .
this model finds its origin in the tachocline where strong shearing motions occur in the solar internal plasma , store and deform the magnetic field , and can give a semi - empirical model of solar 11-yr cycle and reproduced the well known sunspot butterfly diagrams . at the same time
, there are several other approaches .
polygiannakis & moussas ( 1996 ) assume that the sunspot - cycle - modulated irradiance component may be caused by a convective plasma current , this current can drive a nonlinear rlc oscillator
. this model can describe the shape and the related morphological properties of the solar cycles , and give some reasonable interpretations to the long period inactive maunder minimum .
the closeness between the 11-yr schwabe cycle and the 11.86-yr period of jupiter has been noted for a long time . and
this let us to speculate that planetary synodic period may resonate with solar activity ( wood , 1975 ) .
grandpierre ( 1996 ) proposed that the planetary tides of co - alignments can drive the dynamo mechanism in the solar interior and trigger solar eruptions .
they calculate the co - alignments ( conjunction and opposition ) of the earth , venus , and jupiter and find that their co - alignment periods are in the range of 8.7588 13.625 yr , the averaged value is around 11.2 yr , being very close to the observed schwabe cycle of 11-yr .
however , de jager & versteegh ( 2005 ) compare the accelerations due to planetary tidal force , the sun s motion around the gravitational center of the solar system , and the observed acceleration at the level of tachocline , and find that the latter are by a factor of about 1000 larger than the former two , and assert that the planetary tidal force can not trigger the dynamo mechanism observably . however , in my opinion , as the plasma system is always very brittle , even with a very small perturbation , a variety of instabilities are also very easy to develop , accumulate , and trigger a variation as the solar schabe cycle . hence , at present stage , it is very difficult to confirm which factor is the real driver of solar schwabe cycles .
the most bewildering is the harmonic relationship between the long - term cycles @xmath2 and @xmath3 .
this may imply that solar long - term active cycles have wave s behaviors .
however , so far , we do nt know more natures of this kind of cycles .
the dynamo theory can give some reasonable interpretation to the 11-yr solar schwabe cycle , but it can not present even if a plausible explanation to the 51.5-yr cycle and 103-yr secular cycle so far . in 1943 , alfven assumed that solar large - scale dipole magnetic field has the axis coincide with the rotation axis , and the magnetic disturbance in the interior travels along the field line to the surface at alfven speed .
this magnetic disturbance will excite mhd oscillation and transport to the whole sun , the travel time of the mhd oscillation is about 70 80 yr .
this time scale is neither agree with the 11-yr period , nor with the time scale of the secular cycles .
many people believe that the 11-yr solar schwabe cycle is originated from the tachocline which is a thin shear layer of strong differential rotation motion at the base of the solar convection zone ( dikpati , 2006 ) .
helioseismology suggests that the tachocline in thin of the order of 1% of the solar radius ( christensen - dalsgaard & thompson , 2007 ) .
it is well - known that the kelvin - helmholtz instability is very easy to develop around a layer with strong shear motion . from the work of gilliland ( 1985 ) , we know that the time scale of kelvin - helmholtz instability for the whole sun is @xmath37 yr
. then it is possible to assume that the time scale of kelvin - helmholtz instability around the thin tachocline can be reduced to the order of 100-yr .
rashid , jones , & tobias ( 2008 ) pointed out that the evolution of the hydrodynamic instability of the slow tachocline region will occur on timescale of hundred years .
however , we need much of investigations to confirm this assumption and to understand the long - term solar cycles .
the origin of long - term solar cycles is still a big unsolved problem .
from the above analysis and estimations , we obtain the following conclusions : \(2 ) the solar schwabe cycle 24 is in a vale between the two secular cycles ( g3 , and g4 ) , it will be a relatively weak and long active cycle , which may reach to its apex in about 2012 - 2014 and last for about 12 years ; \(3 ) most solar schwabe cycles are left asymmetric , they have rapidly rising phases and slowly decay phases .
the most intriguing is that most of the solar powerful flares occurred in the decay - phase of the solar schwabe cycle .
\(4 ) the secular cycle is possibly associated with the solar inner large scale motions as well as the dynamo processes can be account for the 11-yr schwabe cycles .
however , there are much of works need to do to understand the mechanism of the secular cycles .
the author would like to thank the referee s valuable comments on this paper .
this work was supported by nsfc grant no .
10733020 , 10873021 , cas - nsfc key project ( grant no .
10778605 ) , and the national basic research program of the most ( grant no .
2006cb806301 ) .
babcock , h.w . : 1961 , _ apj _ * 133 * , 572 .
badalyan , o.g . ,
obridko , v. , & sykora , n.j . : 2001 , _ solar phys _ * 199 * , 421 .
bonev , p.p . ,
penev , k.m . , & sello , s. : 2004 , _ apj _ * 605 * , l81 .
christensen - dalsgaard , j. , & thompson , m.j . : 2007 , in solar tachocline , ed . d.w .
hughes , r. rosner , & n.o . , weiss ( cambridge university press ) .
de jager , c. , & versteegh , g.j.m .
: 2005 , _ solar phys _ * 229 * , 175 .
dikpati , m. : 2006 , _ adv space res _ * 38 * , 839 .
eddy , j.a . : 1983 , _ solar phys _ * 89 * , 195 .
frick , p. , galyagin , d. , hoyt , d.v .
, et al . : 1997 , _ astron astrophys _ * 328 * , 670 .
fu , q.j . ,
ji , h. , qin , z.h . , & et al : 2004 , _ solar phys _ * 222 * , 167 .
gilliland , r.l . : 1985 , _ apj _ * 290 * , 344 .
gleissberg , w. : 1939 , _ observatory _ * 62 * , 158 .
gleissberg , w. : 1971 , _ solar phys _ * 21 * , 240 .
grandpierre , a. : 1996 , _ astrophys .
space sci . _ * 243 * , 393 .
hathaway , d.h . : 2009 , _ space sci rev _ * 144 * , 401 .
hiremath , k.m . : 2006 , _ astron astrophys _ * 452 * , 591 .
hiremath , k.m . : 2008 , _ astrophys space sci _ * 314 * , 45 .
hiremath , k.m . : 2009 , arxiv0909.4420[astro - ph.sr ] .
, & wang , j.l . : 2003 , _ chin j astron astrophys _ * 3 * , 391 .
leighton , r.b . : 1964 , _ apj _ * 140 * , 1547 .
leighton , r.b . : 1969 , _ apj _ * 156 * , 1 .
nagovitsyn , yu a. : 1997 , _ astron lett _ * 23 * , 742 .
otaola , j.aq . , &
zenteno , g. : 1983 , _ solar phys _ * 89 * , 209 .
pesnell , w.d . : 2008 , _ solar phys _ * 252 * , 209 .
polygiannakis , j.m . , & moussas , c.p . : 1996 , _ solar phys _ * 163 * , 193 .
rashid , f.q . ,
jones , c.a .
, & tobias , s.m . : 2008 , _ astron astrophys _ * 488 * , 819 .
schove , d.j . : 1979 , _ solar phys _ * 63 * , 423 .
stix , m. : 2003 , _ solar phys _ * 212 * , 3 .
strong , & saba , j. l.r .
: 2009 , _ adv space res _ * 43 * , 756 .
suess , h.e . : 1980 , _ radiocarbon _ * 20 * , 200 .
waldmeier , m. : 1961 , _ the sunspot activity in the years 1610 - 1960 _ , zurich schulthess and company ag .
wang , j.l . , zong , w.g .
, le , g.m .
, et al : 2009 , _ res .
astron astrophys .
_ * 9*,133 .
wolf , r. : 1862 , _ astro mitt zurich _ , 209 .
wood , r.m . : 1975 , _ nature _ * 255 * , 312 . | based on analysis of the annual averaged relative sunspot number ( asn ) during 1700 2009 , 3 kinds of solar cycles are confirmed : the well - known 11-yr cycle ( schwabe cycle ) , 103-yr secular cycle ( numbered as g1 , g2 , g3 , and g4 , respectively since 1700 ) ; and 51.5-yr cycle . from similarities ,
an extrapolation of forthcoming solar cycles is made , and found that the solar cycle 24 will be a relative long and weak schwabe cycle , which may reach to its apex around 2012 - 2014 in the vale between g3 and g4 .
additionally , most schwabe cycles are asymmetric with rapidly rising - phases and slowly decay - phases .
the comparisons between asn and the annual flare numbers with different goes classes ( c - class , m - class , x - class , and super - flare , here super - flare is defined as @xmath0 x10.0 ) and the annal averaged radio flux at frequency of 2.84 ghz indicate that solar flares have a tendency : the more powerful of the flare , the later it takes place after the onset of the schwabe cycle , and most powerful flares take place in the decay phase of schwabe cycle . some discussions on the origin of solar cycles are presented . |
qpt is performed out by noting how input states , that span the space of @xmath0 , map to output states .
the key insight in what follows is that it is not the input states of @xmath0 that are relevant , rather it is the preparation procedures itself , i.e. , the preparation map @xmath11 . for a @xmath20 dimensional system
there are @xmath21 linearly independent states that span its space . however , there are @xmath22 linearly independent operations ( preparations ) that span the space of preparations .
if we determine the corresponding output states for a set of linearly independent preparations then by linearity we have can predict the output state for any preparation .
let us denote this map as @xmath23-map .
the form of @xmath23-map arises naturally when considering the whole process in physical terms : at the beginning of the experiment @xmath2 is in an unknown ( correlated ) state , @xmath24 .
the system is prepared into a known input state by the preparation procedure @xmath11 , followed by a joint unitary dynamics .
the output is given by tracing over the environmental degrees of freedom : @xmath25 \left({\rho^{\mathcal{se}}}\right ) u^\dag \right].\ ] ] we want a map acting on the preparation map @xmath11 and yielding the output state @xmath12 : @xmath26 .
then @xmath23-map is everything on the right hand side of eq .
( [ output ] ) that is not @xmath11 .
the expression for @xmath23-map in terms of matrix indices is @xmath27 above a sum over repeated indices is implied .
@xmath23-map is a ` super super - operator ' that acts on the super operator @xmath11 .
@xmath23-map is a @xmath28 tensor , which is contracted with a preparation @xmath11 , a @xmath29 tensor , yielding the output state @xmath12 , a @xmath30 matrix . in term of matrix indices ,
the action is as follows : @xmath31 again , a sum over repeated indices is implied . in methods , a full derivation for @xmath23-map in the last equation is given .
see fig . 2 for a graphical illustration of @xmath23-map .
figure 2 : _ quantum process tomography with @xmath23-map .
_ at the beginning of the experiment the system - environment state is correlated .
a preparation is made on the system and the corresponding output state @xmath12 is observed .
this process is described by the completely positive map @xmath23 , which is a function of the initial system - environment state and the unitary dynamics .
the @xmath23-map takes preparations @xmath11 to output states @xmath12 . note
that , in standard quantum process tomography state of @xmath1 is a constant of the process , here it is the initial @xmath2 state that is the constant of the process , i.e. , it is a fixed quantity .
physically , the constancy of @xmath24 means that the experiment should be initialised in the same manner for every run , and then a preparation on @xmath0 can be made .
@xmath23-map contains both @xmath32 and @xmath24 ; however knowing @xmath23 is not sufficient to determine @xmath32 and @xmath24 .
as expected , it should not be possible to determine @xmath32 and @xmath24 through measurements and preparations on the system alone without access to the environment .
conversely , @xmath23-map contains all information necessary to fully determine the output state for any preparation of @xmath0 .
the advantage of dealing with the @xmath23-map is that we have separated the preparation procedure from uncontrollable dynamical elements and the initial conditions .
@xmath23-map contains all of the dynamical information for the system and in the next section we will extract some of this information from the @xmath23-map .
first let us mention some properties of @xmath23-map derived in methods : its action on a mixture of preparations is linear , it preserves trace , it preserves hermiticity , and it is completely positive . in methods we show that @xmath33map can be experimentally determined by making a set of linearly independent preparation of the system .
this is similar to what one has to do in standard qpt . in standard qpt a linearly independent set of states
are fed into the process and the corresponding outcomes are observed . knowing the inputs and the outputs the standard process map is determined .
the difference here is that a linearly independent set of preparations are fed in to the process .
this is of our major result of this paper : we have given a prescription to determine the dynamics of a system in an operational way , i.e. , a mapping from preparations to output states .
the @xmath23-map contains the dynamics of the system before any preparation is made on the system .
it is a function of the initial state @xmath24 state as well as the @xmath2 unitary transformation .
@xmath23-map is a tensor , taking its trace with respect to the indices that belong to the initial state of @xmath0 we can obtain the dynamics of the system as if the initial correlations we absent . using this with the knowledge of the initial state of @xmath0 , in methods
we show that from @xmath23 we can derive another matrix , @xmath34 matrix @xmath35 is fully determinable from @xmath33map and the two are the same when there are no initial correlations .
we will call the difference between @xmath23 and @xmath36 , @xmath37 , the _ memory matrix _ : @xmath38 since @xmath23 contains @xmath24 and @xmath36 contains @xmath39 , the difference between the two is a function of only @xmath5 .
the action of the correlation - memory matrix on a preparation yields @xmath40(\chi^{\mathcal{se } } ) u^\dag],\ ] ] which is the coherence coming into the system from the initial correlations . for non - markovian dynamics
the future state of @xmath0 may depend on the initial @xmath2 correlations .
this is the non - markovian ` memory ' due to the initial @xmath2 correlations and it is a key feature of non - markovian dynamics @xcite .
the correlation - memory matrix is an important result for studying non - markovian systems .
it is an operational way of measuring the information that flows into @xmath0 due to correlations at the time of the preparation . once @xmath33map is determined , we have the full knowledge of the dynamics of @xmath0 that is due to the initial correlations .
the correlation - memory matrix provides quantitative information about the initial correlation and it is more than a witness for initial correlations @xcite . for the special case , when the preparation is chosen to be the identity map , we get pure dynamics of the correlation - memory matrix @xmath41,\ ] ] which is the reduced dynamics of @xmath2 correlations .
this is exactly @xmath42 in eq .
[ cpmap ] . from @xmath23-map
we can determine matrices @xmath36 and @xmath43 .
in turn , from @xmath36 we can get @xmath7 ( see eq .
( [ l - cpmap ] ) ) and from @xmath43 we can get @xmath42 , and together they give us @xmath9 of eq .
( [ ncpmap ] ) , which can be a not - completely positive map .
this gives not - completely positive maps an operational meaning .
@xmath23-map is the result of a quantum process tomography procedure for initially correlated system - environment states .
it is acts on the preparation of the initial state of the system , and only contains dynamical information .
we study the properties of @xmath23-map , showing it to be linear , preserving of trace and hermiticity , and completely positive .
dynamical information about the evolution of the initial correlations can be retrieved from @xmath33map , in the form of the correlation - memory matrix @xmath43 .
@xmath23-map allows us to determine the output state for any preparation of the system , while the correlation - memory matrix @xmath44 provides a quantitative expression for the coherence due to the initial correlations .
an important question is when is @xmath23-map relevant ?
clearly , when @xmath0 and @xmath1 are initially uncorrelated then @xmath44 will be zero .
alternatively , just the presence of initial @xmath2 correlations does not warrant for @xmath23-map .
suppose @xmath10 but @xmath45 = 0 $ ] , then the completely positive map of eq .
( [ cpmap ] ) would suffice to describe the dynamics correctly for any preparation of @xmath0 @xcite .
one downside to @xmath33map is that it requires a lot of resources to construct . in standard quantum process tomography @xmath46 input states
are fed through the process and the corresponding output states are determined . to determine @xmath23 map , @xmath47 preparations are necessary , which is a significant growth over the standard procedure . therefore an efficient way , such as compressed sensing @xcite , to determine this map is desirable .
this should be possible , as determining @xmath23-map is equivalent to carrying out @xmath20 standard quantum process tomography procedures .
another limitation that faces the procedure is the assumption that the preparation acts only on the system and not on the environment .
this assumption is crucial , as we are mapping from the set of preparations on the system to the corresponding output states .
if this assumption fails , then we would need to make a set of preparations that span the space of operations on the combined system - environment space .
however , the environment can be arbitrarily large and we do not have any control over it .
therefore the tools given in this article may not be valid when the preparation affects the environment directly .
when the preparation procedure acts on @xmath1 as well as @xmath0 , the positivity of @xmath23-map may be affected .
note that , as long the effect of all preparations on @xmath1 is a constant for then our prescription remains valid .
lastly , since @xmath23-map contains all dynamical information , we are able to construct @xmath7 of eq .
( [ cpmap ] ) from it .
similarly , from the correlation - memory matrix , we can construct @xmath42 of eq .
( [ cpmap ] ) . knowing the two we can determine @xmath9 of eq .
( [ ncpmap ] ) , which can be a not - completely positive map .
this gives operational meaning to not - completely positive dynamical map as the descriptor for the dynamics of the system when identity preparation is made . on the other hand ,
the non - completely positive map is not experimentally determinable without determining @xmath23-map .
finally , it remains an open question , when @xmath48 , is @xmath9 not completely positive ?
the calculations in this sections are done in terms of matrix indices as @xmath23-map and the correlation - memory matrix @xmath43 are nontrivial tensors .
we use the einstein summation notation , i.e. , repeated indices are summed over .
bipartite state of @xmath2 is expressed with four indices with the latin indices belong to @xmath0 and greek indices to @xmath1 .
for instance , the state in eq .
( [ corrmatrix ] ) has the form @xmath49 .
a map acting on a density matrix is written as @xmath50 , where @xmath51 are the sudarshan - kraus operators ( see @xcite ) .
@xmath52 is the complex conjugation of @xmath53 and @xmath54 is the transpose ; together they give hermitian conjugation .
let us rewrite the generalised process equation , eq .
( [ output ] ) , in terms of matrix indices @xmath55 where the sum over @xmath56 is the trace with respect to the environment .
we are interested in the reduced dynamics of @xmath0 as a function of the preparation procedures .
thus , we can pull the preparation map out of everything else and regard it all as a map acting on the preparation map : @xmath57 \left ( { \mathcal{a}}_{r'r'';s 's '' } \right)\\ & = & \label{rawprocessequation } { \mathcal{m}}_{rr'r'';ss 's '' } \left ( { \mathcal{a}}_{r'r'';s 's '' } \right).\end{aligned}\ ] ] in the last equation , the matrix @xmath23 is defined as : @xmath58 let @xmath59 be a set of pure states that linearly span the space of @xmath0 .
there are @xmath21 such matrices .
that is , any state of @xmath0 can be written as a linear sum of these pure states : @xmath60 .
a preparation map acting on @xmath0 is a @xmath29 hermitian matrix .
therefore , any matrix in this space can be spanned by a tensor product of the basis matrices @xmath61 , which is a basis in for @xmath29 space of maps .
there are @xmath22 elements in the basis @xmath61 .
we can write action of one of these basis element on a density operator on @xmath0 as @xmath62 it is crucial to note here that @xmath63 , as these vectors are eigenvectors of the basis elements @xmath64 that do not commute .
these preparations are can be thought of as a projection followed by a rotation .
action of any map on space of @xmath0 acting on the @xmath2 state can be expresses as a linear sum @xmath65\\ & = & \sum_{mn } \alpha^{{(mn)}}p^{{(m)}}p^{{(n)}}\otimes \rho^{{\mathcal{e}}|{{(m)}}},\end{aligned}\ ] ] where @xmath66 are the coefficients that determine @xmath11 in terms of @xmath67 .
@xmath16 is the conditional state of the @xmath1 and @xmath68 $ ] is the probability for the outcome @xmath15 .
knowing the output states corresponding to each of these inputs , @xmath69,\ ] ] along with the success probabilities @xmath70 , for all @xmath71 , is enough to predict the output state for any preparation : @xmath72\\ & = & \sum_{mn } \alpha^{{(mn)}}p^{{(m)}}{{\rm tr}_\mathcal{e}}[up^{{(n)}}\otimes \rho^{{\mathcal{e}}|{{(m ) } } } u^\dag]\\ & = & \sum_{mn } \alpha^{{(mn)}}p^{{(m)}}q^{{(mn)}}.\end{aligned}\ ] ] @xmath23-map can be determined choosing @xmath73 , followed determining the corresponding @xmath74 and @xmath70 , and standard inversion techniques @xcite .
note that any other set of linearly independent preparation can be linearly mapped to the preparations given in eq .
( [ linpreps ] ) , and therefore will suffice . determining all @xmath74 is done by quantum state tomography .
this is equivalent to carrying out @xmath20 standard qpt procedures , one each @xmath16 . additionally measuring @xmath70 is equivalent to doing quantum state tomography of @xmath75 . before moving on a simple example may be useful . for one qubit , we may take the following projectors as a linearly independent basis : @xmath76 note that , this is a linear but not a convex decomposition : @xmath77 .
the eigenvectors of @xmath78 , @xmath79 , @xmath80 , and @xmath81 are @xmath82 , @xmath83 , @xmath84 , and @xmath85 respectively . using these eigenvectors we can write basis elements for the maps that operate on the space of one qubit .
for instance , @xmath86 and so on .
the initial state of the system is labeled by indices @xmath87 and @xmath88 .
tracing over everything else we can find the initial state of @xmath0 ( before preparation ) from @xmath23-map : @xmath89 this is , of course , attainable by doing state tomography at the beginning of the experiment , by measuring the values of @xmath70 from last section . next ,
let us the trace over the system indices @xmath87 and @xmath88 @xmath90 the last equation is exactly the dynamical map in the absence of initial correlations , given in eq .
( [ cpmap ] ) . in other words , in the absence of initial correlations , qpt would yield this map .
this means , even though the @xmath23-map contains the information about uncorrelated @xmath2 state and the correlations separately .
consider the following matrix composed of the matrices in eqs .
( [ initialstates ] ) and ( [ processfromm ] ) @xmath91 the last equation is similar to the expression for the @xmath23-map , except the state of the system and the state of the environment are uncorrelated . writing the state of @xmath2 in @xmath23-map in terms of eq .
( [ corrmatrix ] ) , we get @xmath92 now we can define the correlation - memory matrix as @xmath93 mathematically , @xmath33map acts on the preparation map just as the dynamical map acts on a density operator .
in fact , we are not varying the initial state of the system , rather the preparation procedure on that state .
therefore the linearity of quantum mechanics is preserved for the @xmath23-map acting on different preparation procedures , i.e. @xmath94 = \alpha_1 { \mathcal{m}}{\mathcal{a}}^{(1 ) } + \alpha_2 { \mathcal{m}}{\mathcal{a}}^{(2)}.\ ] ] this is very much like the dynamical maps action on mixtures of states . furthermore ,
if we show that the @xmath33map preserves trace , hermiticity , and positivity on its domain then all of these properties will be preserved on the state space . in other words for any preparation , @xmath95 that preserves trace , hermiticity , and positivity , the action of the @xmath33map on it will yield an output state , @xmath96 , that is unit - trace , hermitian and positive .
let us start with the trace of @xmath23 with respect to the final indices @xmath97 with @xmath98 : @xmath99 & = & \delta_{rs } { \mathcal{m}}_{rr'r'';ss's''}\nonumber\\ & = & u_{r \epsilon , r ' \alpha } { \rho^{\mathcal{se}}}_{r''\alpha , s''\beta}u^*_{r \epsilon , s ' \beta}.\end{aligned}\ ] ] since @xmath100 , then @xmath101 = \delta_{r 's ' } \delta_{\alpha \beta } { \rho^{\mathcal{se}}}_{r''\alpha , s''\beta } = \mathbb{i } \otimes { \rho^{\mathcal{s}}}.\ ] ] a preparation acting on the above matrix will yield @xmath101({\mathcal{a}})={\mbox{tr}}[{\mathcal{a}}({\rho^{\mathcal{s}}})]=1.\ ] ] the implication being @xmath23 preserves the trace of @xmath102 .
as long as the preparation is trace a preserving operation we get a unit - trace matrix for the output state . as with the case of general quantum operations
, matrix @xmath23 is hermitian .
this is easy to see by taking the complex conjugate of matrix @xmath23 , @xmath103 the complex conjugate of @xmath23 is not only the transpose of @xmath23 , but each element of @xmath23 is also transposed .
hence @xmath23 is a hermitian matrix .
the @xmath23-map is composed of a unitary matrix operating on a density matrix .
then we can take the square root of the density matrix to get @xmath104 where @xmath105 and @xmath106 .
we have written the @xmath23-map in operator sum representation , hence it is completely positive .
where @xmath107 are the sudarshan - kraus operators @xcite .
this means , the @xmath23-map acting on any preparation procedure will lead to a physical state .
this was not the case when a standard qpt procedure is carried out on initially correlated @xmath2 states .
the action of @xmath23-map can now be written as @xmath108 the properties shown above are precisely the conditions for a generic quantum operation to preserve trace , hermiticity and positivity .
therefore @xmath33map preserves the attributes on the preparations , which in return will preserve these attributes on the states .
10 kossakowski , a. on quantum statistical mechanics of non - hamiltonian systems . , 247274 ( 1972 ) .
lindblad , g. on the generators of quantum dynamical semigroups .
, 119130 ( 1975 ) .
gorini , v. , kossakowski , a. & sudarshan , e. c. g. completely positive dynamical semigroups of n - level systems .
, 821825 ( 1976 ) .
engel , g. et al .
evidence for wavelike energy transfer through quantum coherence in photosynthetic systems .
, 782786 ( 2007 ) .
meiera , c. & tannor , d. j. non - markovian evolution of the density operator in the presence of strong laser fields .
, 33653376 ( 1999 ) .
geva , e. , rosenman , e. & tannor , d. j. on the second - order corrections to the quantum canonical equilibrium density matrix . , 13801390 ( 2000 ) .
rodrguez - rosario , c. a. & sudarshan , e. c. g. non - markovian open quantum systems .
, 16171634 ( 2011 ) . wolf , m. m. , eisert , j. , cubitt , t. s. & cirac , j. i. assessing non - markovian dynamics .
, 150402 ( 2008 ) .
breuer , h .-
p . , laine , e .-
& piilo , j. measure for the degree of non - markovian behavior of quantum processes in open systems .
, 210401 ( 2009 ) .
rivas , a. , huelga , s. f. & plenio , m. b. entanglement and non - markovianity of quantum evolutions .
, 050403 ( 2010 ) .
devi , a. r. u. , rajagopl , a. k. & sudha open system quantum dynamics with correlated initial states , not completely positive maps and non - markovianity . , 022109 ( 2011 ) .
laine , e .-
m . , piilo , j. & breuer , h .-
witness for initial system - environment correlations in open system dynamics .
, 60010 ( 2010 ) .
smirne , a. , brivio , d. , cialdi , s. , vacchini , b. & paris , m. g. a. experimental investigation of initial system - environment correlations via trace distance evolution .
, 032112 ( 2011 ) .
rossatto , d. z. , werlang , t. , castelano , l. k. , villas - boas , c. j. , & fanchini , f. f. purity as a witness for initial system - environment correlations in open - system dynamics . , 042113 ( 2011 ) .
sudarshan , e. c. g. , mathews , p. m. & rau , j stochastic dynamics of quantum - mechanical systems .
, 920924 ( 1961 ) .
kraus , k. .
springer - verlag , ( 1983 ) .
carteret , h. , terno , d. & zyczkoski , k. dynamics beyond completely positive maps : some properties and applications .
, 042113 ( 2008 ) .
modi , k. , rodrguez - rosario , c. a. & aspuru - guzik , a positivity in the presence of initial system - environment correlation . : 1203.5209 , ( 2012 ) .
shaji , a. & sudarshan , e. c. g. who s afraid of not completely positive maps ? , 4854 ( 2005 ) .
chuang , i. l. & nielsen , m. a. prescription for experimental determination of the dynamics of a quantum black box .
, 24552467 ( 1997 ) .
poyatos , j. f. , cirac , j. i. & zoller , p. complete characterization of a quantum process : the two - bit quantum gate .
, 390393 ( 1997 ) .
nielsen , m. a. , knill , e. & laflamme , r. complete quantum teleportation using nuclear magnetic resonance .
, 5255 ( 1998 ) .
childs , a. m. , chuang , i. l. & leung , d. w. realization of quantum process tomography in nmr .
, 012314 ( 2001 ) .
mitchell , m. w. , ellenor , c. w. , schneider , s. & steinberg , a. m. diagnosis , prescription , and prognosis of a bell - state filter by quantum process tomography bell - state filter by quantum process tomography .
, 120402 ( 2003 ) .
weinstein , y. s. et al .
quantum process tomography of the quantum fourier transform .
, 61176133 ( 2004 ) .
obrien , j. l. et al .
quantum process tomography of a controlled - not gate .
, 080502 ( 2004 ) .
neeley , m. et al .
process tomography of quantum memory in a josephson - phase qubit coupled to a two - level state .
, 523526 ( 2008 ) .
chow , j. m. et al .
randomized benchmarking and process tomography for gate errors in a solid - state qubit . , 090502 ( 2009 ) .
howard , m. et al .
quantum process tomography and lindblad estimation of a solid - state qubit . , 33 ( 2006 ) .
myrskog , s. h. , fox , j. k. , mitchell , m. w. & steinberg , a. m. quantum process tomography on vibrational states of atoms in an optical lattice .
, 013615 ( 2005 ) .
pomyalov , a. , meier , c. & tannor , d. j. the importance of initial correlations in rate dynamics : a consistent non - markovian master equation approach .
, 98108 ( 2010 ) .
yuen - zhou , j. & aspuru - guzik , a. quantum process tomography of excitonic dimers from two - dimensional electronic spectroscopy .
i. general theory and application to homodimers .
, 134505 ( 2011 ) .
yuen - zhou , j. , krich , j. , mohseni , m. & aspuru - guzik , a. quantum state and process tomography of energy transfer systems via ultrafast spectroscopy . , 17615 ( 2011 ) .
kuah , a .- m . , modi , k. , rodrguez - rosario , c. a. & sudarshan , e. c. g. how state preparation can affect a quantum experiment : quantum process tomography for open systems .
, 042113 ( 2007 ) .
modi , k. & sudarshan , e. c. g. the role of preparation in quantum process tomography .
, 052119 ( 2010 ) .
modi , k. preparation of states in open quantum mechanics .
, 253 ( 2011 ) .
nielsen , m. a. & chuang , i. l. .
cambridge university press , ( 2000 ) .
modi , k. .
phd thesis , the university of texas at austin , ( 2008 ) .
[ _ arxiv_:0903.2724 ] emerson , j. et al . symmetrised characterisation of noisy quantum processes . ,
1893 ( 2007 ) .
flammia , s. t. , gross , d. , liu , y .- k . & eisert , j. quantum tomography via compressed sensing : error bounds , sample complexity , and efficient estimators . : 1205.2300 , ( 2012 ) .
we are grateful to a. brodutch , a. rezakhani , c. a. rodrguez - rosario , and keith burnett for valuable conversations .
we acknowledge the financial support of john templeton foundation , national research foundation , and ministry of education in singapore .
part of the work presented here was done while the author was at the university of texas at austin . | a central aim of physics is to describe the dynamics of physical systems .
schrdinger s equation does this for isolated quantum systems .
describing the time evolution of a quantum system that interacts with its environment , in its most general form , has proved to be difficult because the dynamics is dependent on the state of the environment and the correlations with it . for discrete processes , such as quantum gates or chemical reactions
, quantum process tomography provides the complete description of the dynamics , provided that the initial states of the system and the environment are independent of each other .
however , many physical systems are correlated with the environment at the beginning of the experiment . here
, we give a prescription of quantum process tomography that yields the complete description of the dynamics of the system even when the initial correlations are present .
surprisingly , our method also gives quantitative expressions for the initial correlation .
= 1 there is a rich history to the studies of decoherence of quantum systems due to the interactions with the surrounding degrees of freedom . when the dynamics of the system ( @xmath0 ) is markovian it can be described by a master equation @xcite . nowadays
many researchers are interested in systems that are non - markovian , as there is mounting evidence that some natural systems of importance may be non - markovian @xcite and such features may allow to manipulate and control quantum systems in desired ways .
there is also a great deal of interest in systems that are initially correlated with their environments ( @xmath1 ) because non - markovianity and initial system - environment ( @xmath2 ) correlations are intimately related @xcite .
grasping the mathematical and physical aspects of non - markovian systems , especially with initial @xmath2 correlations , has proved to be a tough road .
nevertheless , there is a great deal of progress on deciding whether a system is non - markovian in the recent years @xcite .
however , avoiding the initial @xmath2 correlations is not always possible in reality @xcite .
working with initial correlations in practice has proved to be much trickier than in theory .
this is because the presence of correlations do not allow for a clear definition of the state @xmath0 independent from the state of @xmath1 and vice versa .
physical systems are complicated and have many additional degrees of freedom that are not of experimental interest . yet
these extra degrees of freedom interact with the degrees of interest leading to correlations .
therefore initially uncorrelated @xmath2 state is often an approximation . in theory of open quantum systems ,
discrete quantum transformations are described by the dynamical map formalism @xcite : @xmath3 .
the dynamical map can be thought of as coming from the contraction of @xmath2 unitary dynamics .
let us write the state of @xmath2 as @xmath4 where @xmath5 is the correlations matrix @xcite .
the dynamical map is the mapping from the initial states of @xmath0 to the final states of @xmath0 , resulting from unitary dynamics of the @xmath2 state @xmath6\label{ncpmap}\\ & = & { { \rm tr}_\mathcal{e}}\left[u { \rho^{\mathcal{s}}}\otimes { \rho^{\mathcal{e}}}u^\dag\right]+ { { \rm tr}_\mathcal{e}}\left[u \chi^{\mathcal{se}}u^\dag\right]\\ & = & { \mathcal{b}}^{\rm cp } ( { \rho^{\mathcal{s } } } ) + { \mathcal{b}}^{\rm aff},\label{cpmap}\end{aligned}\ ] ] where @xmath7 is a completely positive map and @xmath8 is the affine correction term due to the initial @xmath2 correlations .
this means that @xmath9 may not a be completely positive map when @xmath10 , nevertheless it fully describes the dynamics of @xmath0 @xcite .
however , to determine such a map experimentally would require preparing different states of @xmath0 while keeping the @xmath2 correlations fixed .
such preparations are not operationally feasible because altering the state of @xmath0 will also alter the @xmath2 correlations .
therefore , a nonpositive dynamical map is not an operationally meaningful quantity .
the operational approach to quantum dynamics relies on the fact that quantum theory is a theory of preparations and measurements .
the experimental method to determine a dynamical map corresponding to a quantum process is called _ quantum process tomography _ ( qpt ) @xcite .
it is the central tool in determining a discrete quantum process ; e.g. quantum gates @xcite or chemical reactions @xcite . to see the difference between qpt and dynamical maps
let us review the four basic steps necessary to carried out qpt @xcite : * input states that span the space of @xmath0 are prepared .
* the input states are sent through the process . *
the corresponding output states are determined by quantum state tomography . *
the knowledge of input states , the corresponding output states , and assuming linearity completely determines the process .
figure 1 : _ standard quantum process tomography .
_ at the beginning of the experiment the system - environment state is uncorrelated . a preparation ( @xmath11 )
is made on the system and the corresponding output state @xmath12 is observed .
this process is described by the completely positive map of eq .
( [ cpmap ] ) , which is a function of initial state of environment and the unitary dynamics .
it maps the initial states of the system to output states @xmath12 .
let us denote input states as @xmath13 and output states as @xmath12 .
the first step of qpt is state preparation .
a preparation procedure takes an unknown state of @xmath0 to a known state of @xmath0 .
mathematically , it is described by a completely positive map acting on the system @xcite .
for instance , consider a set of preparations that project @xmath0 into pure states : @xmath14 \left ( { \rho^{\mathcal{se}}}\right)= p^{{(m)}}\otimes \rho^{{\mathcal{e}}|{{(m)}}}$ ] . since @xmath15 is a pure state , the post - preparation @xmath2 state is fully uncorrelated , where @xmath16 is the conditional state of @xmath1 .
@xmath17 is the identity operator acting on @xmath1 , as we assume that the preparation procedure only acts on @xmath0 and not @xmath1 .
we will discuss the implications of relaxing this assumption in discussions .
lastly , if the preparation is not trace preserving , it should be divided @xmath18 $ ] for normalisation . the @xmath2 evolution , after the preparation yields the output state : @xmath19({\rho^{\mathcal{se } } } )
u^\dag\right ] \label{preq}\\ & = & { { \rm tr}_\mathcal{e}}\left[u p^{{(m)}}\otimes \rho^{{\mathcal{e}}|{{(m ) } } } u^\dag\right ] .
\label{prepaf}\end{aligned}\ ] ] the key difference between the dynamical map in eqs .
( [ ncpmap ] ) and ( [ preq ] ) is the act of state preparation .
because dynamical maps of do accommodate state preparation , they are not operationally defined . in the presence of initial @xmath2 correlations ,
state preparation affects the state of @xmath1 in a nontrivial manner .
that is , the state of @xmath1 in eq .
( [ prepaf ] ) is conditioned by the choice of the preparations . in deriving the standard qpt procedure
it is implicitly assumed that the initial state of @xmath2 is uncorrelated @xcite , i.e. , the state of @xmath1 is thought to be a constant of the problem . when that is the case , the state of @xmath1 in eq .
( [ prepaf ] ) is not conditioned by the preparation procedure . in this case
the derived map for the process is completely positive and is the same as the completely positive dynamical map in eq .
( [ cpmap ] ) .
see fig . 1 for a graphical illustration .
in the presence on initial @xmath2 correlation , the conditional state of @xmath1 will be different for each preparation , and the assumption of linearity in step ( iv ) of qpt is violated , i.e. the map is a function of the preparation procedure .
such maps are nonpositive , nonlinear , or simply put nonsensical @xcite .
it then begs the question , can we determine the dynamics of a system that is initially correlated with @xmath1 ?
this is an important question for two reasons : first , there may be physical system of interest that may have initial correlations .
is it possible to study their dynamics ?
second , for foundational reasons we may care to know what are the limitations in describing the dynamics of physical systems .
a partial solution to these questions was given in @xcite .
in this article we show that not only complete dynamics of initially correlated system can be determined , we can also determine the contribution due to initial correlations . |
the suggestion of a resonating valence bond ground state in spin liquid systems together with possible exotic , fractional spinon excitations has triggered an intense research interest .
@xcite one of the most promising systems for studying these phenomena is the @xmath0 heisenberg antiferromagnet on the kagome lattice .
the ground states proposed by theory for this system include a valence bond crystal @xcite or a resonating valence bond with a gapped @xcite or gapless @xcite excitation spectrum . in several theoretical studies
the raman response of spin liquids on the kagome lattice has been investigated .
the spectrum can consist of low energy modes as well as high energy , broad continua , depending on details of the systems .
@xcite the anisotropy of these signals with respect to the orientation of the electric fields of incident and scattered light can be used to probe local symmetry breaking of the spin degrees of freedom .
@xcite up to now , experimental realizations of the kagome lattice antiferromagnet are plagued by impurities , interlayer coupling , and/or anisotropic magnetic coupling strengths .
it is therefore of importance to understand the influence of these shortcomings on the system s ground state . in both title compounds volborthite ( cu@xmath1v@xmath2o@xmath3(oh)@xmath2@xmath42h@xmath2o ) and vesignieite ( bacu@xmath1v@xmath2o@xmath5(oh)@xmath2 ) the kagome planes
are slightly distorted , resulting in two inequivalent cu sites .
such a reduced symmetry can either enhance or result from a spin gap , from dimerization , or from some kind of long range order . for volborthite
, this distortion leads to two different magnetic exchange interactions @xmath6 and @xmath7 , with an anisotropy of about 20@xmath8 and an average value of @xmath9 k , as determined from magnetic susceptibility and specific heat measurements .
@xcite in magnetic susceptibility measurements a broad maximum is observed around 22 k that points to the onset of short - range magnetic order .
neutron scattering experiments detect nearest neighbor correlations settling for temperatures below @xmath10 k. @xcite finally , @xmath11v - nmr reveals a magnetic transition around @xmath12 k , with anomalously slow energy spin fluctuations persisting down to lowest temperatures .
@xcite the kagome planes in vesignieite are less distorted , with a difference in bond length of only 0.07@xmath8 . @xcite the magnetic exchange interaction is determined as @xmath13 k. bulk susceptibility measurements reveal a broad maximum around 21 k. @xcite the estimated dzyaloshinsky - moriya ( dm ) interaction is around 0.14 @xmath14 , i.e. larger than the quantum critical value of 0.1 @xmath14 @xcite and thus the ground state of vesignieite should be in the nel - ordered regime .
@xcite indeed , @xmath11v nmr shows a partial spin freezing at @xmath15 k. @xcite so far , herbertsmithite ( zncu@xmath1(oh)@xmath16cl@xmath2 ) is considered to be the closest experimental realization of a structurally perfect @xmath0 heisenberg antiferromagnet on a kagome lattice .
@xcite however , this compound shows antisite disorder with up to 10@xmath8 of the magnetic cu@xmath17 ions exchanged by zn@xmath17 ions .
@xcite thereby , a weak interlayer magnetic coupling as well as magnetic vacancies are induced in the kagome layers .
this might alter the system s magnetic properties from those of a perfect kagome lattice .
this possible drawback is avoided in both volborthite and vesignieite , as their intermediate layer is comprised of v@xmath18 ions .
thus , antisite disorder of the cu ions is prevented . in this brief report
, we present inelastic light scattering ( raman scattering ) experiments on powder samples of volborthite and vesignieite .
raman scattering is a sensitive probe of lattice and electronic / magnetic degrees of freedom . in particular ,
raman scattering has been highlighted as a powerful tool to distinguish possible ground states with weak order or local order parameters .
@xcite recently , it has been successfully applied to characterize the excitation spectrum of the spin liquid state in herbertsmithite .
powder samples of the title compounds were prepared as described previously .
@xcite raman scattering experiments have been performed in quasi backscattering geometry using a frequency - doubled @xmath19 nm nd : yag laser with a fixed laser power of 0.5 mw .
such a small power level is essential to avoid an overheating of the powder samples with reduced thermal conductivity .
the samples were installed into a he - cooled closed cycle cryostat with a temperature range of 3 300 k. the spectra were collected via a dilor - xy 500 triple spectrometer by a liquid nitrogen cooled horiba jobin yvon ccd ( spectrum one ccd-3000 v ) .
both compounds volborthite and vesignieite belong to the monoclinic space group @xmath20 .
a symmetry analysis has been carried out .
it yields the following raman active modes for volborthite : @xmath21 and for vesignieite : @xmath22 .
all presented measurements are performed in ( @xmath23 ) polarization , i.e. with parallel incoming and scattered light .
the crystallites in the powder sample are semi - transparent and partially scramble the light polarization upon scattering .
polarization dependent experiments not shown here did not lead to additional information .
in fact , the spectra obtained in crossed ( @xmath24 ) polarization are identical to the ( @xmath23 ) spectra .
both compounds show modes around 800 @xmath25 and 3400 @xmath25 , due to o - h vibrations .
the energy of these modes is much larger than the magnetic exchange interaction .
therefore interactions do not exist and they will not be further discussed here .
[ figure1 ] 3 , 50 , 100 and 200 k in ( @xmath23 ) polarization ; the spectra are shifted in intensity for clarity .
b ) d ) plots of frequency , linewidth and intensity of the phonon mode at 103 @xmath25 ( see arrow ) as function of temperature .
the inset zooms into the low energy region for spectra obtained at 200 and 3 k.,title="fig:",height=250 ] in figure 1 ( a ) raman spectra of volborthite are shown at different temperatures , ranging from 200 to 3 k. with decreasing temperature , the intensity of the phonons increases . furthermore , in the low energy range below @xmath26 @xmath25 , a set of new , sharp phonons appears ( see inset ) .
the temperature evolution of one phonon at 103 @xmath25 is showcased in figure 1 .
the energy scale of this phonon group is in the range of 1@xmath27 3@xmath27 .
we observe a dramatic gain of intensity for temperatures below 50 k. around this temperature , the onset of nearest neighbor spin - spin correlations is stated from magnetic susceptibility . at low temperatures ,
the total number of phonons exceeds the expected 27 modes .
it is suggested that via spin - phonon coupling a structural distortion is gradually taking place in volborthite with decreasing temperature . through this moderate distortion
, frustration is releaved .
this results in an increase in raman active phonons that are energetically in the vicinity of @xmath27 .
[ figure2 ] 3 , 50 , 100 and 250 k in ( @xmath23 ) polarization ; the spectra are shifted in intensity for clarity .
( d ) plots of the temperature development of the broad mode at 115 @xmath25 ( see arrow ) .
the inset shows the spectral region of magnetic scattering for the temperature range 50 3 k.,title="fig:",height=250 ] figure 2 ( a ) plots the raman spectra of vesignieite at different temperatures , in analogy to figure 1 ( a ) .
the phonon modes sharpen and gain in intensity with decreasing temperature , comparable to the spectral changes in volborthite . in the low energy range
a broader mode appears , extending from about 95 @xmath25 to 140 @xmath25 , or roughly 2.6@xmath28
3.8@xmath28 ( see inset ) .
its parameters are plotted in figure 2 ( b ) ( d ) .
this mode has a different , non - lorentzian line shape compared to the phonons .
following the temperature development of the mode s parameters , it becomes pronounced below about 25 k , which is close to @xmath29 k , where short - range magnetic correlations set in . in the structurally perfect kagome lattice compound herbertsmithite ,
a broad scattering continuum is observed at low temperatures , ranging from lowest energies up to @xmath30 .
@xcite this continuum is attributed to spinon excitations from the resonating valence bond ground state .
the mode in vesignieite is less broad , with an onset at a finite energy .
its maximum position around 115 @xmath25 corresponds to @xmath31 , with @xmath13 k. @xcite this is in very good agreement with the estimation of a typical energy for magnetic raman scattering in a heisenberg antiferromagnet @xcite at @xmath32 @xmath25 , where @xmath33 is the number of nearest neighbors and @xmath0 is the spin .
therefore , we attribute this signal to a magnetic scattering process .
the analogous estimation for volborthite leads to @xmath34 @xmath25 . in this energy range
no corresponding signal is observed .
it may be masked by phonons . while neither of the two systems display long range order at finite temperatures , the observation of a spinon continuum comparable to herbertsmithite is absent in both compounds .
therefore , we can conclude that even a slight anisotropy between @xmath6 and @xmath7 of 0.07@xmath8 as in vesignieite has a stronger influence on the spin liquid ground state than antisite disorder in herbertsmithite of about 10@xmath8 .
in summary , we investigated the lattice and magnetic excitations in kagome systems .
our results reveal spin - phonon coupling in the moderately distorted volborthite that lowers the crystal symmetry and enhances spin - spin correlations for temperatures below 50 k. vesignieite , with a close to perfect kagome lattice structure , shows no additional low temperature phonon modes but instead magnetic scattering .
our study demonstrates that the spin dynamics of spin liquid systems depend decisively on the underlying structure .
we would like to thank v. p. gnezdilov for valuable discussions .
this work was supported by dfg , b - igsm and the nth school `` contacts in nanosystems '' . | despite the conceptional importance of realizing spin liquids in solid states only few compounds are known . on the other side the effect of lattice distortions and anisotropies on the magnetic exchange topology and the fluctuation spectrum
are an interesting problem .
we compare the excitation spectra of the two @xmath0 kagome lattice compounds volborthite and vesignieite using raman scattering .
we demonstrate that even small modifications of the crystal structure may have a huge effect on the phonon spectrum and low temperature properties . |
the general theory of relativity is thought to provide the correct classical description for the interaction of gravity with matter .
this description is embodied in a system of partial differential equations on a four dimensional manifold @xmath0 , the so - called _ einstein equations _ , which relate the ricci curvature @xmath1 of an unknown metric @xmath2 to the energy - momentum tensor @xmath3 of matter : @xmath4 to complete the classical picture of a physics based on a collection of fields satisfying a closed system of equations , one must also consider the laws which govern the evolution of the matter fields generating the energy - momentum tensor on the right hand side of @xmath5 .
( one important special case is when there is no matter , the so - called vacuum .
then @xmath5 with vanishing right hand side is a closed system of quasilinear hyperbolic equations . ) in general , one arrives at quite complicated systems of equations .
however , from the perspective of classical physics , all phenomena are in principle described by the solutions of such a system . moreover , for these systems , the initial value problem is natural , just as in classical dynamics .
thus , from one point of view , the general theory of relativity is a classical physical theory that can be studied mathematically in parallel with other field theories of nineteenth - century classical physics .
indeed , the equations of general relativity exhibit similar local behavior with other equations of evolution as regards , for instance , the issues of local existence and uniqueness of solutions to the initial value problem .
when one turns , however , to the initial value problem in the large , the einstein equations present features that have no analogue in other typical equations of mathematical physics .
the subject of this talk will be what appears , at least at first sight , as the most pathological of these features , namely the possibility of loss of uniqueness of the solution of the initial value problem _ without loss of regularity .
_ this possibility is at the center of what is known as the _ strong cosmic censorship conjecture _ formulated by penrose @xcite .
the reason why the theory of the initial value problem in the large for the einstein equations is richer than for other non - linear wave equations is that the global geometry of the characteristics is not constrained _ a priori _ by any other structure .
this geometry , which corresponds precisely to the conformal geometry for the vacuum equations , is _ a priori _ unknown .
it turns out that many features of the initial value problem for hyperbolic equations that one takes for granted actually depend on certain global properties of the geometry of the characteristics ; the question of uniqueness indicated above is one of these .
the best way to gain some intuition for what kind of conformal geometric structure develops in the course of evolution in general relativity and what are the implications of this structure is to carefully examine the special solutions of the theory .
in fact , almost all conjectures and intuition regarding the theory in the end derives from simple properties of such solutions . moreover
, since our focus of interest is global _ geometric _ structure , there is no substitute in building intuition than a good pictorial representation .
this talk will rely very much on such `` pictures '' .
it should be noted , however , that in the spherically symmetric context in which we shall be working , these `` pictures '' , besides conveying intuition , also carry complete and precise information and can be treated on the same level as symbols or formulas .
the assumption of spherical symmetry and associated pictorial representations will be carefully discussed in the next section .
we will then proceed to examine a series of special solutions which will lead to a particular initial value problem .
finally , theorems describing the solutions of the initial value problem will be formulated and their proofs will be discussed . in regard to uniqueness , it turns out that there is always a spacetime which can be uniquely associated to initial data - manifold @xmath6 , along with a symmetric @xmath7-tensor @xmath8 satisfying the constraint equations that would arise if @xmath8 were to be the second fundamental form of @xmath0 realized as a hypersurface in a ricci flat @xmath9-manifold . ] .
this is the so - called _ maximal domain of development _
it is the `` biggest '' spacetime which admits the given initial hypersurface as initial data and is at the same time _ globally hyperbolic _
, i.e. all inextendible causal curves intersect the initial hypersurface precisely once .
this latter property ensures that the domain of dependence property holds .
the question of uniqueness in general relativity is thus the issue of the _ extendiblity _ of this maximal domain of development .
if it is extendible , then the solution is not unique .
since , as noted earlier , there is really no substitute for a pictorial representation , we defer further discussion of this till later on .
the current state of affairs in the theory of quasilinear hyperbolic partial differential equations in several space variables is such that global , large data problems appear beyond reach . for there to be any hope of making headway
, it seems that some sort of reduction must be made to a problem where the number of the independent variables is no more than two . for hyperbolic equations of evolution ,
such reductions in general are accomplished by considering symmetric solutions , or equivalently , symmetric initial data . in general relativity ,
symmetry assumptions are formulated in terms of a group which acts by isometry on the spacetime and preserves all matter .
the only @xmath7-dimensional symmetry group that is compatible with the notion of an isolated gravitating system , i.e. that can act on asymptotically flat spacetimes , is @xmath10 .
solutions invariant under such an action are called _ spherically symmetric_. as we shall see below , most of the expected phenomena of gravitational collapse of isolated gravitating systems , and the fundamental questions that these phenomena pose , can be suggested by the spherically symmetric solutions of various einstein - matter systems .
moreover , the conformal structure of these solutions , which is the essential ingredient for the phenomena we wish to discuss , can be completely represented on the blackboard .
( or on paper ! ) the reason for this is simple : the space of group orbits @xmath11 can be given the structure of a @xmath7-dimensional lorentzian manifold .
restricting to @xmath12 which are _ maximal domains of development _ of initial data , it follows that these can be globally conformally represented as bounded domains in @xmath13-dimensional minkowski space
. the images of such representations are called _ penrose diagrams _ ; from these , the conformal geometry can be immediately read off as the characteristics are just the lines at @xmath14 or @xmath15 radians from the horizontal : @xmath16
to gain some familiarity with these diagrams , it is perhaps best to begin with @xmath9-dimensional minkowski space from this point of view , i.e. with the penrose diagram of the maximal domain of development of minkowski initial data .
here the diagram is as follows : @xmath17 the @xmath18 referred to above is a function on @xmath12 defined to be a multiple of the square root of the area of the group orbit corresponding to the points of @xmath12 .
the line labelled @xmath19 is thus the axis of symmetry .
the line labelled future null infinity " is not part of the spacetime but should be thought of as a `` boundary '' at infinity .
the same applies to its two endpoints , `` spacelike infinity '' and `` future timelike infinity '' .
the latter corresponds to the `` endpoint '' of all inextendible future timelike geodesics .
the above minkowski space is of course future causally geodesically complete , i.e. all causal curves can be extended to infinite affine parameter .
thus we have the analogue of global existence and uniqueness . that these properties of minkowski space are stable to small perturbations ( _ without _ any symmetry assumptions )
is a deep theorem of christodoulou and klainerman @xcite .
having understood the conformal diagram of minkowski space , we turn to a more interesting solution : the schwarzschild solution .
this is actually a one - parameter family of solutions ( the parameter is called mass and denoted by @xmath20 ) which contains minkowski space ( the case where @xmath21 ) .
as it is a spherically symmetric vacuum solution , its non - triviality in the case @xmath22 must be generated by topology .
any cauchy hypersurface has two asymptotically flat ends and topology @xmath23 .
`` downstairs '' , this corresponds to a line with two @xmath24 endpoints , not intersecting an axis of symmetry . for convenience , we will choose a time - symmetric initial hypersurface .
the maximal development of this `` schwarzschild '' initial data then looks like : @xmath25 the point @xmath26 depicted where @xmath27 is a minimal surface `` upstairs '' in the initial hypersurface .
as the initial hypersurface was chosen to be time symmetric , this minimal surface is also what is called marginally trapped .
the `` ingoing '' and `` outgoing '' null cones emanating from this surface , which correspond to the two null rays through @xmath26 `` downstairs '' , can thus not reach `` future null infinity '' .
( this is related to the so - called `` singularity theorems '' of penrose . )
moreover , all timelike geodesics emanating from the point @xmath26 reach in finite time the curve @xmath19 .
this curve is not an axis of symmetry but a @xmath28 singularity !
that is to say , there is no extension of the spacetime above through @xmath19 with a continuous lorentzian metric .
this singular behavior of the schwarzschild solution may at first appear to be an undesirable feature .
indeed , historically , it was first considered exactly as such .
but it turns out in fact that the behavior outlined above would provide the `` ideal '' scenario for the end - state of gravitational collapse .
this has to do with two specific features of this solution : 1 .
labelling the null rays emanating from the minimal surface as @xmath29 , it turns out that there is a class of timelike observers , namely those who do not cross @xmath29 , who can observe for infinite time and whose causal past is completely regular .
the singularity is hidden inside a black hole , and @xmath29 is called the event horizon . to be more precise about the `` completeness '' property of the region outside the black hole ,
fix some outgoing null geodesic which intersects future null infinity and parallel translate a conjugate null vector ( i.e. a null vector in the other direction ) .
the affine length of the null curves joining this null geodesic and @xmath29 , where the affine parameter is determined by the aforementioned vector , goes to @xmath30 .
in particular , future null infinity can be thought to have infinite affine length .
one says that the solution pocesses a `` past complete future null infinity '' or a complete domain of outer communications . 2 .
the above spacetime is future inextendible as a @xmath28 metric .
thus , according to the discussion in the introduction , this means that the schwarzschild solution is unique even in the class of very low regularity solutions .
the significance of this fact will become clear later .
at this point it should be noted that while the schwarzschild solution indeed provides intuition about black holes , it can not give insight as to whether these can occur in evolution of data where no trapped surfaces are present initially , i.e. whether the kind of behavior outlined above is related in any way to the endstate of gravitational collapse . that properties 1 and 2 above are indeed general properties of solutions was proven by christodoulou @xcite for the spherically - symmetric einstein scalar field equations : @xmath31 @xmath32 @xmath33 for generic solutions of the initial value problem , the penrose diagram obtained by christodoulou is as follows : @xmath34 in @xcite , however , christodoulou explicitly constructs solutions with conformal diagram : @xmath35 these are so - called `` naked singularities '' .
one might ask why should one consider the coupling with the scalar field . the natural case to consider first , it would seem , is the vacuum . a classical theorem of birkhoff , however , states that the only spherically symmetric vacuum solutions are schwarzschild .
thus , matter _ must _ be included to give the problem enough dynamical degrees of freedom in spherical symmetry .
the scalar field is in some sense the simplest , most natural choice .
, etc ] as far as property 1 is concerned , christodoulou s results are the best evidence yet that this is indeed a property of `` realistic '' graviational collapse .
it turns out however that there is another `` competing '' set of evidence that indicates that the behavior of christodoulou s solutions related to property 2 does not represent `` realistic collapse '' .
( remember that property 2 is the central question for us , as this is what determines the notion of uniqueness . )
this evidence is provided again by the intuition given by special solutions .
one may consider the schwarzschild family of solutions as embedded in a larger , 2-parameter family of solutions called the _
kerr solutions_. here the parameters are called mass and angular momentum , and schwarzschild corresponds to vanishing angular momentum . for all non - vanishing values of angular momentum ,
the internal structure of the black hole is completely different , and , as we shall see momentarily , much more `` problematic '' , as property 2 will fail .
thus the introduction of even an arbitrarily small amount of angular momentum a phenomenon that can not be `` seen '' by spherically - symmetric models
seems to change everything and cast doubt on the conclusions derived from the spherically symmetric einstein - scalar field model . to summarize our `` unhappy '' situation
, it seems that the phenomenon which plays a fundamental role in the issue we want to study is incompatible with the assumptions we have to make in order to render it mathematically tractable .
it would seem that understanding the black hole region of realistic gravitational collapse using a spherically symmetric model is a lost cause .
fortunately , there is a `` solution '' to this problem ! the effect of angular momentum on gravity turns out to be similar to the effect of charge .
( as john wheeler puts it , charge is a poor man s angular momentum . ) indeed , there is a very close similarity between the conformal structure of the kerr family and a 2-parameter _ spherically symmetric _ family of solutions to the einstein - maxwell equations : @xmath36 @xmath37 @xmath38}=0,\ ] ] @xmath39 the so - called _ reissner - nordstrm _ solution . here
the parameters are mass @xmath20 and charge @xmath40 . for @xmath41
one retrieves the schwarzschild family , while for @xmath42 one obtains : @xmath43 the @xmath19 singularity of the schwarzschild solution has disappeared ! the above spacetime is completely regular up to the edges .
these new edges that `` complete '' the triangle , however , are at a _
finite _ distance from the initial data , in the sense that all timelike geodesics joining those edges with the initial hypersurface have finite length .
this solution thus has a regular future boundary and is extendible ( in @xmath44 ! ) beyond it .
what fails at the boundary of this maximal domain of development of initial data is thus not the regularity of the solution , but rather , global hyperbolicity .
any extension of @xmath12 will contain past inextendible causal geodesics not intersecting the intial hypersurface : @xmath45 points in such an extension but not in @xmath12 itself can not be determined by initial data , in the exact same way that the solution to a linear wave equation @xmath46 at the point @xmath47 depicted below , can not be uniquely determined by its values in the shaded set @xmath48 : @xmath49 what has happened in the reissner - nordstrm solution is that the situation depicted above developed , but from an @xmath48 that was complete .
the physical interpretation of the above situation is that the classical principle of determinism fails , but without any sort of loss in regularity that would indicate that the domain of the classical theory has been exited .
it is in that sense that this kind of behavior is widely considered by physicists to be problematic . on the other hand , numerical calculations ( penrose and simpson @xcite ) on the behavior of linear equations on a reissner - nordstrm background indicate that a naturally defined derivative blows up at the cauchy horizon .
this was termed the blue - shift effect .
thus , penrose argued , the pathological behavior of the reissner - nordstrm solution might be unstable to perturbation . this led him to conjecture , more generally , _ strong cosmic censorship _ for generic initial data , in an appropriate class , the maximal domain of development is inextendible . in view of our discussion in the introduction , this can be thought of as the conjecture that , for generic initial data , the solution is unique wherever it can be defined . of course
, the context in which this should be applied to ( i.e. what equations , what class of initial data should be considered , etc ) and the notion of extendibility are left open .
we will comment more on that later .
it might seem at first that the proper setting for discussing the problem of whether cauchy horizons arise in evolution , for generic data in spherical symmetry , is the einstein - maxwell equations .
unfortunately , these suffer in fact from the same drawback as the einstein vacuum equations , namely they do not possess the required dynamical degrees of freedom .
one necessarily has to include more matter , and again the simplest choice , as in the work of christodoulou , is a scalar field .
thus one is easily led to the coupled einstein - maxwell - scalar field system : @xmath50 @xmath37 @xmath38}=0,\ ] ] @xmath32 @xmath51 @xmath52 it turns out that in spherical symmetry the maxwell part of the equation decouples . since the scalar field @xmath53 carries no charge , a non - trivial maxwell field
can only be present if an initial complete spacelike hypersurface has non - trivial topology .
in particular , this model is not suitable for considering the formation of black holes , as in the work of christodoulou .
thus we will consider the problem where there is already a black hole present initially . to take the simplest possible formulation that captures the essense of the problem at hand
, one can prescribe initial data for the system on two null rays , such that one corresponds to the event horizon of a reissner - nordstrm solution , and the other carries `` arbitrary '' matching data : @xmath54 to make the most of the method of characteristics , we introduce null coordinates , i.e. coordinates @xmath55 such that the metric on @xmath12 takes the form @xmath56 . here
we select the @xmath57-axis to be the event horizon , and the @xmath58 axis to be the conjugate ray on which we presribe our data .
the unknowns are then just @xmath18 , @xmath59 and @xmath53 , and the electromagnetic part contributes a constant @xmath60 which is computed from the initial data .
to write the equations as a first order system , we define @xmath61 , @xmath62 , @xmath63 , @xmath64 , and also @xmath65 , what one can call the `` renormalized '' hawking mass is defined to be @xmath66 .
the renormalized version has the property that it is constant in the reissner - nordsrm solution and coincides with @xmath20 at future null infinity .
] , defined by @xmath67 we then have : @xmath68 @xmath69 @xmath70 @xmath71 @xmath72 @xmath73 @xmath74
we can now state the theorems of @xcite . on the one hand we have : after restricting the range of the @xmath58 coordinate , the penrose diagram of the solution of the i.v.p .
described above is as follows : @xmath75 moreover , @xmath18 extends to a function on the cauchy horizon with the property @xmath76 as @xmath77 ( where @xmath78 is the constant value of @xmath65 on the initial _ event _ horizon ) , and the metric can be continuously extended globally across the cauchy horizon .
on the other hand , we have : for `` generic '' initial data in the class of allowed initial data , @xmath65 blows up identically on the cauchy horizon .
in particular , the solution is inextendible across the cauchy horizon as a @xmath79 metric .
thus , strong cosmic censorship is true , according to theorem 2 , if formulated with respect to extendibility in @xmath79 or higher , but false , according to theorem 1 , if formulated with respect to extendibility in @xmath28 ( see @xcite for reasons why one might want to require @xmath28 . ) in any case , the formation of a null `` weak '' singularity indicates a qualitatively different picture of the internal structure of the black hole from any of the previous models described above , and from the original expectations of penrose .
the scenario of theorems 1 and 2 was first suggested by israel and poisson @xcite who put forth some heuristic arguments .
subsequently , a large class of numerics was done for precisely the equations considered here ( see @xcite for a survey ) .
because of the blow - up in the mass , this phenomenon was termed `` mass inflation '' .
it should be noted that the imposition of reissner - nordstrm data on the event horizon is somewhat unnatural if the data is viewed as having arisen from generic data for a characteristic value problem where the @xmath57-axis is in the domain of outer communications .
similar results to the above theorems , however , can in fact be proven for a wide class of data which includes the kind conjectured to arise from the aforementioned problem .
these results will appear in @xcite .
the relavence of such an extension will only become clear , however , if the problem of determining the correct `` generic '' decay on the event horizon is mathematically resolved .
our initial data are trapped , i.e. @xmath80 and @xmath81 are both nonpositive .
these signs are then preserved in evolution .
note that from the equation @xmath82 it follows that @xmath83 is non - increasing in @xmath58 .
what gives the analysis of our equations _ in the black hole region _
its characteristic flavor is the fact that @xmath84 is potentially infinite when integrated for constant @xmath58 along the whole range of @xmath57 ( it is indeed infinite in initial data , i.e. @xmath85 ) , and that this infinity can appear in the equations ( for instance in @xmath86 ) with either a positive or negative sign , depending on the sign of @xmath87 .
note that , by contrast , in the domain of outer communications , this infinity is killed by the @xmath88 term since @xmath89 on outgoing rays . in the domain of development of our initial data , @xmath18 is bounded above by its initial constant value on the event horizon @xmath90 , in view of the signs of @xmath81 and @xmath80 . in the reissner - nordstrm solution
, the sign of @xmath87 goes from negative near the event horizon to positive near the cauchy horizon , while @xmath91 remains constant in @xmath58 and thus infinite , accounting for both what is called the infinite red shift near the event horizon ( this makes objects crossing the event horizon slowly disappear to outside observers as they are shifted to the red ) and the infinite blue shift near the cauchy horizon ( this accounts for the instability of the cauchy horizon to _ linear _ perturbations ) . for general solutions of the initial value problem , some of these features of the sign of @xmath87 turn out to be stable , while others do not .
in particular , theorems 1 and 2 together imply that the sign must become negative near the cauchy horizon , and not positive ! to attack this initial value problem ,
it is clear that the behavior of this sign is the first thing that must be understood .
it turns out that before the effects of the linear instability start to play a role , three geometrically distinct regions develop in evolution , a red - shift , no - shift , and stable blue - shift region , characterized by @xmath92 respectively : @xmath93 in the red - shift region , @xmath84 is unbounded as @xmath77 , but it appears with a `` favorable '' sign ( favorable as far as controlling @xmath65 is concerned bigger , and there is an extra @xmath80 on the denominator of @xmath94 . ] ) , in the no - shift region , @xmath95 is uniformly bounded , and in the stable blue - shift region , @xmath84 grows with @xmath58 but at a rate `` less '' than the growth of certain natural derivative of @xmath53 .
these facts allow us to control all quantities reasonably well up until the future boundary of the stable blue - shift region , though completely different arguments must be applied to each subsequent region as it develops in evolution from the previous one .
of course , all this work seems only to have pushed forward the problem from the original initial segments to the future boundary of the `` stable blue - shift region '' .
but in fact our new `` initial '' conditions on the future boundary of the stable blue - shift region are much more favorable .
the stable blue - shift region that has preceded it ensures that @xmath80 has a sufficiently fast decay rate in @xmath58 .
( remember that blue - shift regions tend to make @xmath96 smaller , so they are favorable for controlling @xmath18 , but unfavorable for controlling @xmath65 . )
once this rate can be shown to be preserved , it follows by integrating @xmath58 that one can bound @xmath18 _ a priori _ away from @xmath97 in its future , and thus prove the existence of the solution up to the cauchy horizon .
it is clear from what we have said above that if the unstable region remains a blue - shift region ( see left diagram below ) , there is no problem .
( this is of course what happens for the reissner - nordsrm solution itself . )
@xmath98 the danger is if a new red - shift region develops ( see right diagram above ) .
it turns out , however , that a new _ a priori _
estimate @xmath99 is available in this case , which is independent of the size of @xmath65 .
this makes use of the fact that there is a @xmath65 hidden also in the denominator in @xmath100 .
( the estimate depends , however , on our knowledge of the new initial condition of @xmath18 and a particular bootstrap assumption on its future behavior ; in particular this estimate does not hold in the original `` red - shift '' region . )
this is the last element in the proof of theorem 1 .
we now discuss the proof of theorem 2 .
as has been noted , at the level of perturbation theory @xcite , one does see an instability , caused by the blue - shift region .
( requiring some derivative of the scalar field to be positive initially , @xmath101 will decay in @xmath58 , and thus @xmath102 will decay in @xmath57 , at a slower rate than the decay of @xmath81 in @xmath57 , so that the natural derivative @xmath103 . ) on the other hand , the non - linearity tends to diminish this effect , since if the mass indeed increases , in view of theorem 1 , we have a reappearance of a red - shift region ( see diagram on the right above ) .
as remarked earlier , this tends to make @xmath96 ( and also @xmath104 ) bigger , and thus @xmath105 smaller .
the proof must thus encorporate something beyond `` linear theory arguments '' .
this extra ingredient is supplied by a very powerful monotonicity peculiar to black hole interiors , or more specifically , to `` trapped regions '' .
integration of @xmath106 and @xmath107 then implies that if @xmath102 and @xmath101 are initially of the same signs coordinate .
this condition and the non - vanishing of a particular derivative of @xmath53 at the origin together define the `` generic '' class of initial data to which theorem 2 applies .
] , let s say non - negative , then they remain non - negative , and in fact @xmath108 and @xmath109 . now
integration of @xmath110 and using that @xmath111 is also non - increasing in @xmath58 , yields that for @xmath112 , @xmath113 , @xmath114 it should be mentioned that in view of the sign of @xmath81 , both sides of the above inequality are positive .
the broad outline of the proof of theorem 2 is as follows : first assume that the spacetime looks very much like reissner - nordstrm .
then the linear theory more or less applies , and applying the bounds for @xmath102 in the equation @xmath110 , one obtains @xmath115 , which is a contradiction .
thus one is reduced to proving that any spacetime `` quantitatively different '' from reissner - nordstrm must have @xmath115 .
it is not possible here to explain precisely what `` quantitatively different '' has to mean . to give a taste of the kind of arguments involved in this `` non - linear part '' of the proof , we will be content to show that assuming only that @xmath65 is bounded below by a positive number plus its reissner - nordstrm value ( this is indeed a quantitative difference ) @xmath116 it follows from @xmath117 that @xmath65 must in fact blow up identically on the cauchy horizon
. if @xmath118 is the future boundary of the stable blue - shift region , the fact that @xmath119 implies that given any @xmath120 , a sequence of points @xmath121 can be constructed so that the mass differences @xmath122 are all greater than @xmath123 : @xmath124 but in view of @xmath117 , this mass difference can be added on @xmath125 to yield infinite mass at the point where @xmath125 meets the cauchy horizon .
suffice it to say that the non - linear analysis of black hole regions is quite different than the analysis we are used to .
which parts of this spherically - symmetric picture generalize and which do not remains to be seen .
piotr chruciel _ on the uniqueness in the large of solutions of the einstein s equations ( `` strong cosmic censorship '' ) _ australian national university , centre for mathematics and its applications , canberra , 1991 | this talk will describe some recent results @xcite regarding the problem of uniqueness in the large ( also known as _
strong cosmic censorship _ ) for the initial value problem in general relativity .
the interest in the issue of uniqueness in this context stems from its relation to the validity of the principle of determinism in classical physics . as will be clear from below , this problem does not really have an analogue in other equations of evolution typically studied .
moreover , in order to isolate the essential analytic features of the problem from the complicated setting of gravitational collapse in which it arises , some familiarity with the conformal properties of certain celebrated special solutions of the theory of relativity will have to be developed .
this talk is an attempt to present precisely these features to an audience of non - specialists , in a way which hopefully will fully motivate a certain characteristic initial value problem for the spherically - symmetric einstein - maxwell - scalar field system .
the considerations outlined here leading to this particular initial value problem are well known in the physics relativity community , where the problem of uniqueness has been studied heuristically @xcite and numerically @xcite . in @xcite , the global behavior of generic solutions to this ivp , and in particular , the issue of uniqueness ,
is completely understood .
only a sketch of the ideas of the proof is provided here , but the reader may refer to @xcite for details . |
in lattice numerical studies gauge invariant quantities are usually computed . on the other hand , gauge covariant quantities also provide important information .
the well known examples are quark and gluon propagators , requiring complete gauge fixing , monopoles and p - vortices , which study needs only partial gauge fixing .
the first lattice calculations of the gluon propagator were performed in landau gauge @xcite . nowadays these results are significantly improved , also various gauges are used ( see e.g. @xcite ) .
maximal abelian gauge ( mag ) , used to demonstrate the dual super conductor confinement mechanism , is especially interesting .
propagators in this gauge were not explored carefully enough so far .
the first such study of propagators in the coordinate space was performed in ref .
there was no study of propagators in the momentum space . in this paper
our aim is to close this gap .
we present our results of the high statistics calculation of propagators of the diagonal and the off - diagonal gluon in @xmath0 lattice gauge theory in mag .
complete gauge fixing is achieved by using abelian landau gauge to fix remaining abelian gauge degrees of freedom .
we use the standard parameterization of @xmath0 link matrices , @xmath2 .
then gauge fields are defined as follows : @xmath3 we call @xmath4 the diagonal gluon field , and @xmath5 the off - diagonal gluon field .
the maximal abelian gauge condition in a differential form is @xmath6 nonperturbative fixing of this gauge amounts to the minimization of the functional @xmath7 = \int d^4 x ( ( a^1_\mu(x))^2 + ( a^2_\mu(x))^2).\ ] ] in our simulations the simulated annealing algorithm @xcite with 20 randomly generated gauge copies is employed to minimize the effect of gribov copies .
after mag , only @xmath1 degrees of freedom remain unfixed .
we fix them using @xmath1 landau gauge .
the differential lattice gauge condition we are using is @xmath8 this condition implies that the diagonal field @xmath9 is not transversal at a finite lattice spacing .
another gauge condition @xmath10 is not considered here and will be discussed elsewhere @xcite .
condition ( [ eq : contland ] ) is equivalent to the maximization of the functional @xmath11 = \sum_{x,\mu } \cos\theta_\mu(x),\ ] ] using only @xmath1 gauge transformations .
a local maximization algorithm with 30 random gauge copies is used to accomplish this task . as the `` stop criterion '' for the algorithm we use @xmath12_{new } - f[\theta]_{old}}{f[\theta]_{new } } < \varepsilon = 10^{-8}.\ ] ]
we calculate the diagonal propagator @xmath13 and the off - diagonal propagator @xmath14 where fourier transform @xmath15 is defined as @xmath16 @xmath17 the physical lattice momenta @xmath18 are related to @xmath19 as follows : @xmath20 since both diagonal and off - diagonal fields are not transversal the general structure of diagonal and off - diagonal propagators is @xmath21 thus we have four structure functions @xmath22 , which are really not independent .
note that in nonabelian landau gauge , @xmath23 , there would be only one formfactor : @xmath24
to calculate propagators the lattice with @xmath25 ( 138 configurations ) at @xmath26 were simulated .
the lattice spacing is @xmath27 at this @xmath28 .
the behavior of transversal parts of propagators in the ultraviolet region is the same as in nonabelian landau gauge as our gauge corresponds to that gauge in the limit of large momenta .
both gluons are well described by the perturbative formula @xmath29 , see fig .
[ fig : both ] .
the longitudinal part @xmath30 tends to zero as @xmath31 tends to infinity . and
@xmath32 is small in comparison with other structure functions ( see inset in fig . [
fig : both ] ) .
it increases linearly with increasing @xmath31 .
similar result was obtained in @xcite for the longitudinal part of the photon propagator in 3d compact qed when gauge condition was chosen allowing nonzero longitudinal part at the finite lattice spacing as in our case .
indications that the longitudinal part tends to zero in the continuum limit were found in @xcite .
the sharp increase of @xmath33 at low momenta seems to be related to imprecise @xmath1 landau gauge fixing .
we repeated computations on our smaller @xmath34 lattices ( 300 gauge field configurations ) with reinforced condition ( [ eq : stopcrit ] ) : @xmath35 decreased down to @xmath36 .
we found that @xmath37 , @xcite . in ir region one can see a very strong suppression of @xmath38 in comparison with @xmath39 demonstrating the essence of the abelian dominance @xcite , fig.[fig : both ] . at the same time
@xmath40 approaches @xmath38 at @xmath41 .
the off - diagonal propagator thus becomes : @xmath42 fit to this form at @xmath41 with @xmath43 gives @xmath44 . in the range
@xmath41 the best fit ( @xmath45 ) to @xmath39 is given by the formula @xmath46 we obtain @xmath47 .
this shows that behavior of @xmath39 in the infrared region is qualitatively similar to that of the propagator in nonabelian landau gauge @xcite .
our results clearly show that off - diagonal gluons are suppressed at low momenta thus providing the explanation of the abelian dominance established in numerical studies of mag .
effects of the finite volume and incomplete gauge fixing should be further investigated .
the authors are grateful to f.v .
gubarev , m.n .
chernodub , a. schiller , g. schierholz and v.i .
zakharov for useful discussions .
m. i. p is partially supported by grants rfbr 02 - 02 - 17308 , rfbr 01 - 02 - 117456 , rfbr 00 - 15 - 96 - 786 , intas-00 - 00111 , and crdf award rpi-2364-mo-02 .
s. m. is partially supported by grants rfbr 02 - 02 - 17308 and crdf mo-011 - 0 .
is supported by jsps fellowship .
f. d. bonnet , p. o. bowman , d. b. leinweber , a. g. williams and j. m zanotti , phys .
d * 64 * ( 2001 ) 034501 .
+ c. alexandrou , p. de forcrand and e. follana , phys .
d * 63 * ( 2001 ) 094504 .
k. amemiya and h. suganuma , phys .
d * 60 * ( 1999 ) 114509 .
g. s. bali , v. bornyakov , m. muller - preussker and k. schilling , phys .
d * 54 * ( 1996 ) 2863 .
v. g. bornyakov , s. m. morozov , m. i. polikarpov , to be published .
m. n. chernodub , e. m. ilgenfritz and a. schiller , phys .
* 88 * ( 2002 ) 231601 ; m. n. chernodub , e. m. ilgenfritz and a. schiller , arxiv : hep - lat/0208013 .
t. suzuki and i. yotsuyanagi , phys . rev .
d * 42 * ( 1990 ) 4257 . | we study propagators of diagonal and off - diagonal gluons in the momentum space in maximal abelian gauge of @xmath0 lattice gauge theory .
remaining @xmath1 degrees of freedom are fixed using landau gauge .
we find substantial difference between the propagator of the diagonal and the off - diagonal gluon in the infrared region .
the propagator of the off - diagonal gluon is suppressed in comparison with that of the diagonal gluon at small momenta . in the ultraviolet region
both propagators behave as in nonabelian landau gauge . |
homogeneous coordinate rings of grassmannians are among the most important examples of cluster algebras .
cluster structures in these rings @xcite play a prominent role in applications of cluster theory arising in connection with integrable systems , algebraic lie theory , poisson geometry , teichmller theory , total positivity , and beyond ; see , e.g. , @xcite and references therein . within cluster algebra theory proper ,
grassmannians provide the most concrete and accessible examples of naturally defined cluster algebras of infinite mutation type . despite their importance
, cluster structures on grassmannians are not well understood at all , apart from a few special cases .
just a tiny subset of their cluster variables have been explicitly described ; we do not know which quivers appear in their seeds ; we do not understand the structure of their underlying cluster complexes ; and so on .
let @xmath3 denote the grassmann manifold of @xmath4-subspaces in an @xmath5-dimensional complex vector space .
the corresponding cluster algebra has finite type ( i.e. , has finitely many seeds ) if and only if @xmath6 .
all of the problems mentioned above are open for any grassmannian of infinite cluster type , so in particular for @xmath7 , @xmath8 .
( the case @xmath9 has been well understood since the early days of cluster algebras , see ( * ? ? ?
* section 12.2 ) . )
we advocate the point of view that many aspects of cluster structures on grassmannians are best understood within a broader range of examples coming from classical invariant theory .
recall that the homogeneous coordinate ring of @xmath3 ( with respect to a plcker embedding ) is isomorphic to the ring of @xmath0 invariants of @xmath5-tuples of vectors in a @xmath4-dimensional complex vector space @xmath1 .
more general rings of @xmath0 invariants of collections of vectors and linear forms have been thoroughly studied by classical invariant theory .
we conjecture that _ every such ring carries a natural cluster algebra structure_. in fact , there are typically many such structures , depending on a choice of a cyclic ordering of the tuples of vectors and covectors . in this paper , we prove this claim in the special case when @xmath1 is @xmath2-dimensional .
our main result ( theorem [ th : main ] ) describes a family of cluster structures in the rings of @xmath10-invariants of collections of vectors and covectors in @xmath11 .
the cluster structure ( and moreover the mutation type of the corresponding quiver ) depends on the choice of a _
cyclic signature _ , a binary word describing the order in which vectors and covectors are arranged around a circle .
an accurate formulation of this result requires a description of ( some of ) the seeds defining the cluster structure in question ; this in turn relies on the development of fairly technical combinatorial vocabulary , not to mention the requisite background on cluster algebras , tensor calculus , and basic invariant theory .
this makes it impractical to include the precise statement of our main result in this introduction .
what makes the case @xmath7 special , and amenable to combinatorial approaches that can not be straightforwardly generalized to higher dimensions ? from our perspective , the distinguishing characteristic of the @xmath2-dimensional case is the existence of a beautiful _ web basis _ discovered by greg kuperberg @xcite .
our investigations were largely motivated by the desire to understand the cluster - theoretic significance of kuperberg s basis .
the main new feature of our approach is an emphasis on the multiplicative properties of the web basis , which along with its compatibility with tensor contraction play a central role in the study of cluster algebra structures in these rings of invariants . while all of our results hold for arbitrary rings of @xmath10-invariants , with an ( almost ) arbitrary cyclic ordering of vectors and covectors , many of our theorems are new already in the case of grassmannians @xmath12 .
the list of such results includes : * a large family of new `` non - plcker '' cluster variables ; * compatibility of cluster structures across different values of @xmath5 ; * examples of `` imaginary '' elements in the web basis ; * examples of negative structure constants ; * a description of clusters associated with arbitrary triangulations of an @xmath5-gon .
the latter description extends a construction given by j. scott @xcite , and suggests an extension to other riemann surfaces , which we plan to pursue in a separate paper .
we describe conjectural ways in which each cluster structure in a classical ring of @xmath10-invariants is intrinsically determined by the corresponding web basis .
we believe that this type of relationship extends to many other settings in which multiplicative properties of some distinguished additive basis in a given commutative ring dictate a `` canonical '' choice of a cluster structure in the ring .
we also believe that , conversely , each cluster algebra of geometric type has an additive basis with certain remarkable properties ; see conjecture [ conj : a - and - b ] .
perhaps most significantly , we formulate a conjectural combinatorial description ( see conjecture [ conj : cluster - variables - are - trees ] ) of _ all _ cluster variables in each of our cluster algebras . to the best of our knowledge ,
this is the first class of cluster algebras of infinite mutation type for which such a description has been proposed .
we next outline the general plan of the paper , and review the contents of each section .
sections [ sec : tensors][sec : webs ] cover the requisite background , some of which may be familiar to the reader . section [ sec : tensors ] is a reminder on the basic notions of tensor calculus .
section [ sec : rings - of - invariants ] introduces our main object of study , the ring @xmath13 of @xmath0 invariants of collections of @xmath14 vectors and @xmath15 covectors ( i.e. , linear forms ) in a complex vector space @xmath1 .
fundamentals of cluster algebras are reviewed in section [ sec : cluster - algebras ] .
we limit ourselves to the case of cluster algebras defined by quivers , as this is all the generality we need . at the end of the section , we state sufficient conditions that ensure that a particular seed , i.e. , a quiver whose vertices are labeled by elements of a given ring @xmath16 ,
defines a cluster algebra structure in @xmath16 . starting with section [ sec : tensor - diagrams ]
, we assume that the space @xmath1 is @xmath2-dimensional . section [ sec : tensor - diagrams ] is a primer on _ tensor diagrams _ , a particular kind of combinatorial gadgets that can be used to define @xmath0 invariants .
this diagrammatic calculus , popular among physicists , has the advantage of reducing cumbersome calculations with tensors and invariants to repeated applications of certain ( purely combinatorial ) local transformation rules called skein relations .
section [ sec : webs ] presents the crown jewel of the theory of tensor diagrams : kuperberg s construction of the web basis in @xmath13 .
this basis , which depends on a choice of a cyclic signature , consists of _ web invariants _ , the @xmath0 invariants defined by planar tensor diagrams without short internal cycles .
sections [ sec : special - invariants ] and [ sec : special - seeds ] introduce the key combinatorial construction of the paper : a family of `` special seeds '' designed to define a cluster structure in @xmath13 . in section [ sec : special - invariants ]
, we describe the `` special invariants '' that appear in those seeds , and state their basic properties .
we then explain in section [ sec : special - seeds ] how each triangulation of a convex @xmath17-gon gives rise to an extended cluster consisting of @xmath18 special invariants , and to a quiver whose vertices are labeled by them .
an impatient reader unwilling to labor through the rather intricate technicalities of this construction may decide to skip the details , and go directly to section [ sec : main - theorem ] .
sections [ sec : main - theorem][sec : zoo ] present the main results and conjectures of the paper .
( the proofs are deferred until later . )
the main theorem ( theorem [ th : main ] ) asserts that each of the special seeds described in section [ sec : special - seeds ] defines a cluster algebra structure in the ring of invariants @xmath13 .
this cluster structure is independent of the choice of such a seed ; in other words , all special seeds are mutation equivalent .
the cluster structure does however depend on the choice of a signature .
in fact , this choice affects the ( cluster ) type of the resulting cluster algebra , and whether it is of finite or infinite type ( resp . , finite or infinite mutation type ) .
this can be seen in figure [ fig : boundary - signatures-5678 ] that lists the cluster types of @xmath13 for all signatures with @xmath19 .
our main construction has nice functoriality properties : it is preserved by the duality between vectors and covectors , and respected by the natural embeddings of smaller invariant rings into larger ones , induced by the forgetful maps ( dropping a ( co)vector from a collection ) or the maps defined by taking a cross product of two consecutive ( co)vectors . as a corollary , we establish that if a tensor diagram is a planar tree , then the corresponding web invariant is a cluster or coefficient variable . in section
[ sec : main - conjectures ] , we discuss many conjectural connections between our main construction and kuperberg s web basis .
in particular , we expect this basis to contain all cluster monomials .
furthermore , we expect two cluster variables to be compatible if and only if their product is a web invariant .
we formulate criteria that should distinguish cluster variables / monomials among more general web invariants .
while the web basis has many wonderful properties , it _ may _ have negative structure constants ( cf .
conjecture [ conj : strong - positivity ] and proposition [ prop : negative - structure - const ] ) .
this is perhaps not so surprising in light of the discovery , made by m. khovanov and g. kuperberg @xcite , that the web basis is generally different from the _ dual canonical basis _ ,
i.e. , the basis dual to g. lusztig s canonical basis @xcite . putting things into a cluster - theoretic context
allows us to `` explain '' the khovanov - kuperberg smallest counterexample : it corresponds to the square of the simplest web invariant which is not a cluster variable .
as established by b. leclerc @xcite and p. lampe @xcite , for the cluster type at hand the appropriate element of the dual canonical basis is given by the quadratic chebyshev polynomial of the second kind .
using the chebyshev polynomial of the first kind recovers the corresponding element of the `` atomic '' basis of p. sherman and a. zelevinsky @xcite , while the square lies in the _ dual semicanonical basis _ , see @xcite .
it is then natural to ask : does the web basis always coincide with the appropriate dual semicanonical ( or `` generic '' ) basis , in the sense of @xcite ?
section [ sec : arborization ] begins by our favorite conjecture of the paper ( conjecture [ conj : cluster - variables - are - trees ] ) that describes , in simple combinatorial terms , the entire set of cluster monomials in @xmath13 . according to this conjecture ,
a cluster monomial is an @xmath0 invariant that possesses two alternative presentations by a single tensor diagram : first , by a ( planar , non - elliptic ) web ; second , by a ( possibly non - planar ) forest .
if , in addition , this invariant does not factor ( so that this forest is actually a tree ) , then the invariant is a cluster or coefficient variable .
we then present an explicit _ arborization _ algorithm that conjecturally detects whether a given web defines a cluster monomial ( resp .
, a cluster or coefficient variable ) by applying a sequence of skein relations that either transform the web into a forest ( resp . , tree ) , or else determine that this is impossible .
section [ sec : zoo ] presents a gallery of fairly complicated examples of webs illustrating various phenomena that we discovered .
these include : non - arborizable webs , both `` real '' and `` imaginary '' ( in the sense of b. leclerc @xcite ) ; `` fake '' exchange relations involving web invariants ; and the aforementioned negative structure constants . sections [ sec : special - proofs][sec : other - proofs ] contain the proofs of the results stated in sections [ sec : special - invariants][sec : main - theorem ] and [ sec : arborization ] .
specifically , the properties of special invariants and special seeds formulated in sections [ sec : special - invariants][sec : special - seeds ] are proved in sections [ sec : special - proofs][sec : properties - of - special - seeds ] .
the construction of the quiver associated with a special seed defined by an arbitrary triangulation is outlined in section [ sec : building - a - quiver ] .
the main theorem [ th : main ] is proved in section [ sec : proof - main ] .
other results stated in sections [ sec : main - theorem ] and [ sec : arborization ] are proved in section [ sec : other - proofs ] .
the last two sections play the role of appendices .
section [ sec : t - fan ] describes , in precise combinatorial terms , the construction of a special seed associated with a particular choice of a triangulation .
this is in principle sufficient to identify a cluster structure in the ring @xmath13 once the main theorem has been proved .
section [ sec : examples - of - seeds ] shows a couple of additional examples of special seeds , illustrating the main construction .
the paper contains a large number of pictures , which are best viewed in .
we were influenced in our work by the research of many friends and colleagues on the subject of cluster algebras .
we are grateful to arkady berenstein , thomas lam , bernard leclerc , jan schrer , david speyer , and andrei zelevinsky for helpful advice and stimulating discussions .
we thank bernard leclerc and gregg musiker for their comments on the earlier version of the paper .
our main sources of inspiration outside cluster theory included the timeless texts by h. weyl @xcite and v. popov e .
vinberg @xcite on the foundations of classical invariant theory ; the pioneering work by g. kuperberg @xcite on the web bases ; and the groundbreaking paper by v. fock and a. goncharov @xcite on @xmath20-teichmller spaces .
we are grateful to igor dolgachev , rob lazarsfeld , ivan losev , ezra miller , and mircea musta for answering some of our questions on classical invariant theory .
our work on this paper began and ended at the mathematical sciences research institute in berkeley , ca .
it started in 2008 during msri s combinatorial representation theory program , and was completed in 2012 during the cluster algebras program .
we are grateful to msri for an excellent work environment .
the main results of this paper were first reported at the hausdorff institute ( bonn ) and the abel symposium ( balestrand ) in april and june 2011 , respectively .
the first author thanks the organizers of these events for their hospitality and support .
let @xmath1 be a @xmath4-dimensional complex vector space . the word
_ vector _ will always refer to an element of @xmath1 .
the elements of the dual space @xmath21 ( the linear forms on @xmath1 ) are called _
covectors_. a _ tensor _ @xmath22 of _ type _
@xmath23 is an element of the tensor product @xmath24 or alternatively a multilinear map @xmath25 in this paper , we consistently use the latter viewpoint .
we will also need a slight notational variation that allows an arbitrary ordering of the _ contravariant _ arguments of @xmath22 ( vectors in @xmath1 ) and its _ covariant _ arguments ( covectors in @xmath21 ) .
the simplest examples of tensors are vectors ( linear forms on @xmath21 ) and covectors ( linear forms on @xmath1 ) ; they have types @xmath26 and @xmath27 , respectively .
tensors of type @xmath28 are cryptomorphic to linear operators in @xmath29 .
indeed , a tensor @xmath30 corresponds to a linear map @xmath31 that sends @xmath32 to the linear form on @xmath21 given by @xmath33 .
the _ identity tensor _
@xmath34 is the type @xmath28 tensor that corresponds to the identity operator on @xmath1 .
one can accordingly define the _ trace _ of a type @xmath28 tensor , denoted by @xmath35 .
for example , @xmath36 .
the @xmath4th exterior power of @xmath21 is one - dimensional .
let us fix an element in @xmath37 , a _ volume form _ on @xmath1 .
the corresponding _ volume tensor _
@xmath38 has type @xmath39 .
its evaluation @xmath40 is the oriented volume of the parallelotope with sides @xmath41 .
the _ dual volume tensor _
@xmath42 of type @xmath43 is defined by the identity @xmath44 let @xmath22 and @xmath45 be tensors of types @xmath23 and @xmath46 , respectively . the _ tensor product _
@xmath47 is a tensor of type @xmath48 defined by @xmath49 in some instances , the arguments of @xmath22 and @xmath45 are re - shuffled in the process , resulting in different variants of the notion of a tensor product .
we will also need the concept of a _ contraction _ of a tensor with respect to a pair of arguments of opposite variance , i.e. , a vector argument and a covector argument . for a tensor @xmath22 of type @xmath23 given by , its contraction with respect to , say
, the first covariant and the first contravariant arguments is the type @xmath50 tensor @xmath51 obtained by viewing @xmath22 as the function of those two arguments ( temporarily fixing all the rest ) and evaluating the trace of the resulting tensor : @xmath52 one can also define a contraction of two tensors with respect to a contravariant argument of one of them and a covariant argument of another .
the result is obtained by first computing the tensor product of the two tensors , then contracting as above . +
contracting a vector @xmath53 and a covector @xmath54 produces a scalar @xmath55 .
more generally , an @xmath17-fold contraction of a type @xmath23 tensor @xmath22 against an ordered collection of @xmath15 covectors and @xmath14 vectors returns the evaluation of @xmath22 at this collection . +
contracting the two arguments of a type @xmath28 tensor against each other yields the trace of the corresponding linear operator . +
* given two tensors of type @xmath28 , contracting a contravariant argument of one tensor against the covariant argument of the other tensor corresponds to composition of respective linear operators .
two - fold contraction of two such tensors ( matching each contravariant argument to the covariant argument of the other tensor ) yields the trace of the product of the corresponding linear operators . + * 5 . * a @xmath4-fold contraction of the volume tensor against its dual yields the scalar @xmath56 .
all of the above can be recast in the basis - dependent language of coordinates and matrices , once we designate a `` standard '' basis @xmath57 in @xmath1 satisfying @xmath58 ( note that in view of , the dual basis @xmath59 in @xmath21 has the similar property : @xmath60 after a standard basis has been chosen , tensors of type @xmath23 become @xmath17-dimensional arrays of complex scalars ( the _ components _ of a tensor ) indexed by tuples of @xmath15 `` row indices '' and @xmath14 `` column indices ; '' each index takes values from 1 to @xmath4 .
in particular , vectors become column vectors ( with @xmath4 components ) , covectors become row vectors , tensors of type @xmath28 become @xmath61 matrices , and so on .
the trace becomes the usual trace of a matrix .
the evaluation of a volume tensor at a @xmath4-tuple of column vectors is given by the determinant of a square matrix formed by placing them side - by - side in the given order .
the contraction of a tensor with respect to a row index and a column index is obtained by summing over all assignments of matching values to these two indices . in coordinate notation described above
, the identity tensor is given by the identity matrix @xmath62 , also known as the _ kronecker symbol_. in _ einstein notation _
, commonly used in physics , the row indices are placed as superscripts while column indices appear as subscripts ; one then writes the kronecker symbol as @xmath63 . the volume tensor on @xmath1 is given in the coordinates associated with the standard basis as the _ levi - civita symbol _ @xmath64 defined by @xmath65 the dual volume tensor on @xmath21 is given by the same symbol ( denoted @xmath66 in einstein notation ) .
the special linear group @xmath0 naturally acts on both @xmath1 and @xmath21 ; to make the latter a left action , one sets @xmath67 , for @xmath32 , @xmath68 , and @xmath69 .
the group @xmath0 consequently acts on the vector space @xmath70 and therefore on its coordinate ( polynomial ) ring . in this paper , we focus our attention on the ring @xmath71^{\sl(v)}\ ] ] of @xmath0-invariant polynomials on @xmath72 . in coordinate notation , this ring is described as follows .
consider two matrices filled with indeterminates : the matrix @xmath73 of size @xmath74 and the matrix @xmath75 of size @xmath76 .
the ring @xmath13 consists of polynomials in these @xmath77 variables which are invariant under the transformation that simultaneously replaces @xmath78 by @xmath79 , and @xmath80 by @xmath81 , for any matrix @xmath82 .
one example of such polynomial is a _
plcker coordinate _
@xmath83 where @xmath84 is a @xmath4-element subset of columns in @xmath78 ; by definition , @xmath83 is the @xmath61 minor of @xmath78 obtained by selecting the columns in @xmath84 .
( in coordinate - free terms , we evaluate the volume tensor on the @xmath4-element subset , labeled by @xmath84 , of the last @xmath14 factors in . )
one similarly defines the `` dual '' plcker coordinate @xmath85 , for @xmath34 a @xmath4-element subset of rows , as the corresponding maximal minor of @xmath80 .
another example of an @xmath0 invariant is a bilinear form @xmath86 where @xmath87 is a row index for @xmath80 , and @xmath88 a column index for @xmath78 ; thus , this invariant is obtained by pairing a particular @xmath21 factor in with a @xmath1 factor .
the ring @xmath13 is one of the archetypal objects of classical invariant theory ; see , e.g. , ( * ? ? ?
* chapter 2 ) , @xcite , , ( * ? ? ?
* section 11 ) , @xcite .
it was explicitly described by hermann weyl @xcite in the 1930s in terms of generators and relations .
the first fundamental theorem of invariant theory states that the ring @xmath13 is generated by the following multilinear polynomials ( tensors ) : * the @xmath89 ( `` dual '' ) plcker coordinates @xmath85 , * the @xmath90 plcker coordinates @xmath83 , and * the @xmath91 pairings @xmath92 .
thus , in a sense , all @xmath0 invariants are obtained from three fundamental ones : the volume tensors on @xmath1 and @xmath21 , and the identity tensor .
weyl also described the ideal of relations among the generators @xmath85 , @xmath83 , and @xmath92 .
his second fundamental theorem , which we will not rely upon , states that this ideal is generated by the quadratic _ grassmann - plcker relations _ for the @xmath85 s , the similar relations for the @xmath83 s , and the non - homogeneous ( for @xmath93 ) relations @xmath94 ( cf . ) .
classical linear algebraic groups acting in various spaces of tensors .
the central example of this theory at least from our current perspective is the ring of @xmath0 invariants of @xmath95-tuples of vectors in a @xmath4-dimensional complex vector space @xmath1 .
the _ first fundamental theorem of invariant theory _ asserts that this ring is generated by the @xmath96 invariants ... the first fundamental theorem implies that the ring @xmath97 is isomorphic to the homogeneous coordinate ring @xmath98 $ ] of the grassmann manifold @xmath99 of @xmath4-dimensional subspaces in @xmath100 with respect to its plcker embedding ( see , e.g. , ( * ? ? ?
* corollary 2.3 ) ) .
it is not hard to show that @xmath101 is isomorphic to the homogeneous coordinate ring of the two - step partial flag manifold @xmath102 we do not know a similar interpretation of @xmath13 for general values of @xmath15 and @xmath14 . in our discussions of the spaces @xmath13
, it will be important to distinguish between their incarnations that involve different orderings of the contravariant and covariant arguments .
to this end , we will use the notion of a _ signature _ , a word in the alphabet @xmath103 that reflects the order of the @xmath104 arguments of an invariant @xmath105 . specifically ,
each of the @xmath15 factors @xmath21 will be represented by the symbol @xmath106 , and each of the @xmath14 factors @xmath1 by the symbol @xmath107 .
thus for example an @xmath108-invariant polynomial @xmath109 is an invariant of signature @xmath110 $ ] .
we denote by @xmath111 the ring of @xmath0 invariants of signature @xmath112 .
if @xmath112 consists of @xmath15 copies of @xmath106 and @xmath14 copies of @xmath107 , we say that it has _ type _
thus for @xmath112 of type @xmath23 , we have @xmath113 .
the natural action of the @xmath17-dimensional torus defines a multi - grading of the ring @xmath113 .
if an invariant @xmath114 is multi - homogeneous of degrees @xmath115 in its @xmath104 arguments , we then call the tuple @xmath116 the _ multidegree _ of @xmath105 . for example , the plcker coordinate @xmath117 has multidegree @xmath118 ( the first @xmath4 entries are 1 s ) . there are known combinatorial formulas for the dimensions of multigraded components of the ring @xmath13 in terms of the iterated littlewood - richarsdon rule . in some cases , more explicit formulas can be given . for example , the multilinear component of @xmath97 ( i.e. , the space of @xmath119-invariant @xmath14-covariant tensors in @xmath120 ) is only nontrivial when @xmath4 divides @xmath14 ; if it does , then the dimension of this space is equal to the number of standard young tableaux of rectangular shape @xmath121 . as such ,
this dimension is given by the hooklength formula .
we will need the following basic properties of invariant rings .
[ lem : properties - of - inv - rings ] let @xmath122 be a finite - dimensional complex vector space , and @xmath123 a reductive algebraic group .
then the algebra of invariants @xmath124 is a finitely generated normal domain .
if @xmath125 , then @xmath124 is a unique factorization domain .
being a subring of a polynomial ring , @xmath13 is a domain . by ( the modern version of ) hilbert s theorem , it is finitely generated ; as mentioned above , h. weyl @xcite described the generators explicitly .
finally , the ring @xmath13 is factorial : [ lem : properties - of - rabv ] the algebra of invariants @xmath13 is a finitely generated unique factorization domain .
this section offers a quick introduction to cluster algebras , limited in scope to the immediate needs of our project .
further details on cluster algebra fundamentals can be found in @xcite .
the only non - standard material in this section is proposition [ prop : cluster - criterion ] ( and its corollary [ cor : cluster - criterion ] ) which provides a test for verifying that a given set of algebraic / combinatorial data inside a commutative ring makes the latter into a cluster algebra .
cluster algebras are a class of commutative rings endowed with a combinatorial structure of a particular kind .
the concept can be defined in varying degrees of generality . in this paper
, we will only need the notion of a cluster algebra which ( i ) is of _ geometric type _ , ( ii ) has @xmath126 as its field of scalars , and ( iii ) has _ skew - symmetric exchange matrices_. to simplify terminology , we call such gadgets `` cluster algebras '' without further qualifications . a cluster algebra is built inside an ambient field @xmath127 of rational functions in the variables of an initial _ extended cluster _ @xmath128 .
the indeterminates @xmath129 are the initial _
cluster variables _ ; @xmath130 are the _ coefficient _ ( or _ frozen _ ) _ variables_.
the combinatorial data defining a cluster algebra are encoded in a _ quiver _
@xmath131 , a finite oriented loopless graph with no oriented 2-cycles .
some vertices of @xmath131 are designated as _
mutable _ ; the remaining ones are called _
let @xmath132 be a mutable vertex in a quiver @xmath131 .
the _ quiver mutation _
@xmath133 transforms @xmath131 into the new quiver @xmath134 via a sequence of three steps . at the first step , for each pair of directed edges @xmath135 passing through @xmath132 , introduce a new edge @xmath136 ( unless both @xmath78 and @xmath80 are frozen , in which case do nothing ) . at the second step ,
reverse the direction of all edges incident to @xmath132 . at the third step , repeatedly remove oriented 2-cycles until unable to do so .
see figure [ fig : quiver - mutation ] .
( 40,15)(0,4 ) ( 0,0)(0,0)@xmath78 ( 20,0)(0,0)@xmath80 ( 0,20)(0,0)@xmath137 ( 20,20)(0,0)@xmath132 ( 40,20)(0,0)@xmath53 ( 16,0)(-1,0)12 ( 16,20)(-1,0)12 ( 36,20)(-1,0)12 ( 20,16)(0,-1)12 ( 0,4)(0,1)12 ( 3,3)(1,1)14 @xmath138 ( 40,15)(0,4 ) ( 0,0)(0,0)@xmath78 ( 20,0)(0,0)@xmath80 ( 0,20)(0,0)@xmath137 ( 20,20)(0,0)@xmath139 ( 40,20)(0,0)@xmath53 ( 4,20)(1,0)12 ( 24,20)(1,0)12 ( -1.3,4)(0,1)12 ( 1.3,4)(0,1)12 ( 20,4)(0,1)12 ( 17,17)(-1,-1)14 ( 37,17)(-1,-1)14 quiver mutations can be iterated _ ad infinitum _ , using an arbitrary sequence of mutable vertices of an evolving quiver .
this combinatorial dynamics drives the algebraic dynamics of _ seed mutations _ that we describe next .
[ def : seeds ] let @xmath127 be a field containing @xmath126 .
a _ seed _ in @xmath127 is a pair @xmath140 consisting of a quiver @xmath131 as above together with a collection @xmath141 , called an _ extended cluster _ , consisting of algebraically independent ( over @xmath126 ) elements of @xmath127 , one for each vertex of @xmath131 . the elements of @xmath141 associated with the mutable vertices are called _ cluster variables _ ; they form a _ cluster_.
the elements associated with the frozen vertices are called _ frozen variables _ , or _ coefficient variables_. a _ seed mutation _
@xmath133 at a mutable vertex associated with a cluster variable @xmath132 transforms @xmath140 into the seed @xmath142 defined as follows .
the new quiver is @xmath143 the new extended cluster is @xmath144 where the new cluster variable @xmath139 replacing @xmath132 is determined by the _ exchange relation _ @xmath145 ( the two products are over the edges directed at and from @xmath132 , respectively . ) to illustrate , the exchange relation associated with the quiver mutation @xmath133 shown in figure [ fig : quiver - mutation ] is @xmath146 , while applying @xmath147 to the quiver on the right would invoke the exchange relation @xmath148 .
we note that the mutated seed @xmath149 contains the same coefficient variables as the original seed @xmath140 .
it is easy to check that one can recover @xmath140 from @xmath149 by performing a seed mutation at @xmath139 .
definition [ def : seeds ] deviates from the usual convention of using the field of rational functions @xmath150 as the ambient field @xmath127 .
this distinction is inconsequential since , as we shall see , all the action is taking place inside @xmath150 .
[ def : cluster - algebra ] two seeds that can be obtained from each other by a sequence of mutations are called _ mutation - equivalent_. the _ cluster algebra @xmath151 _ associated to a seed @xmath140 is defined as the subring of @xmath127 generated by all elements of all extended clusters of the seeds mutation - equivalent to @xmath140 . thus , to construct a cluster algebra ,
one begins with an arbitrary _ initial seed _
@xmath140 in @xmath127 , repeatedly applies seed mutations in all possible directions , and takes the @xmath126-subalgebra generated by all elements of @xmath127 appearing in all seeds produced by this recursive process .
note that the above construction of @xmath151 does not depend , up to a natural isomorphism , on the choice of the extended cluster @xmath141 : everything , save for the embedding of the resulting cluster algebra into @xmath127 , is determined by the initial quiver @xmath131 , and indeed by its mutation equivalence class .
the _ rank _ of a cluster algebra @xmath151 is the number of cluster variables in each of its seeds , or equivalently the number of mutable vertices in each of its quivers . when a quiver @xmath131 undergoes a mutation , its mutable part ( that is , the induced subquiver on the set of mutable vertices ) does so , too .
thus the mutable parts of the quivers at different seeds of a given cluster algebra @xmath152 are all mutation equivalent .
this mutation equivalence class determines the ( cluster ) _ type _ of @xmath152 .
the same ring can be endowed with cluster structures of different type ; see , e.g. , ( * ? ? ?
* example 12.10 ) . many more examples of this kind appear later in this paper .
a cluster algebra is said to be of _ finite type _ if it has finitely many distinct seeds .
one of the basic structural results of cluster theory is the _ finite type classification _ : [ ] [ th : finite - type ] a cluster algebra is of finite type if and only if it has a quiver whose mutable part is an orientation of a disjoint union of dynkin diagrams .
thus the terminology is consistent : the property of being of finite type depends only on the cluster type of a cluster algebra .
if a cluster algebra @xmath152 ( say of rank @xmath95 ) has a quiver whose mutable part is an orientation of a dynkin diagram @xmath153 , then we say that @xmath152 is `` of type @xmath153 . ''
this convention applies as well to the extended dynkin diagrams of affine and elliptic ( or extended affine ) types ; see , e.g. , ( * ? ? ?
* section 12 ) for further details .
it is worth noting that all orientations of any such diagram @xmath153 are mutation equivalent whenever @xmath153 is a tree ( thus in particular in finite type ) .
furthermore , orientations of non - isomorphic diagrams are mutation inequivalent .
another basic property of cluster algebras is the _
laurent phenomenon _ : [ ] [ th : laurent ] any element of a cluster algebra @xmath151 is expressed in terms of the extended cluster @xmath141 as a laurent polynomial .
furthermore , none of the coefficient variables appears in the denominator of this laurent expression ( reduced to lowest terms ) .
since @xmath151 is generated by cluster variables from the seeds mutation equivalent to @xmath140 , theorem [ th : laurent ] can be restated as saying that each of those cluster variables is given by a laurent expression of the aforementioned kind .
the laurentness part of theorem [ th : laurent ] is a special case of ( * ? ? ?
* theorem 3.1 ) .
the last statement of theorem [ th : laurent ] was established in ( * ? ? ?
* proposition 11.2 ) , modulo a technical condition ( * ? ? ?
* ( 11.3 ) ) which is in fact satisfied in all applications appearing in this paper .
this condition can be removed using ( * ? ? ?
* theorem 3.7 ) in combination with ( * ? ? ? * conjecture 5.4 ) ; the latter was proved ( in the generality considered in the current paper ) in ( * ? ? ?
* theorem 1.7 ) .
many important rings arising in lie theory possess a natural structure of a cluster algebra .
the list includes homogeneous coordinate rings of partial flag manifolds , schubert varieties , and double bruhat cells in semisimple lie groups ; see , e.g. , @xcite . of particular importance to us is the example of a grassmannian @xmath99 . as mentioned earlier , its homogeneous coordinate ring ( with respect to the plcker embedding )
is canonically isomorphic to the ring @xmath154 of @xmath119-invariants of configurations of @xmath14 vectors in a @xmath4-dimensional vector space .
this ring has a natural cluster algebra structure , first defined in j. scott s ph.d .
thesis @xcite ( cf .
also @xcite ) .
it is important to note that this cluster structure on @xmath97 depends on the choice of a cyclic ordering of the @xmath14 vectors ; choosing a different ordering results in a different ( albeit isomorphic ) cluster structure . any cluster algebra
, being a subring of a field , is an integral domain ( and under our conventions , a ) .
conversely , given such a domain @xmath16 , one may be interested in identifying @xmath16 as a cluster algebra . as an ambient field @xmath127 , we can always use the quotient field @xmath155 .
the challenge is to find a seed @xmath140 in @xmath155 such that @xmath156 .
we next present a set of conditions that are sufficient to ensure that a particular choice of a seed solves this problem .
recall that an integral domain @xmath16 is called _ normal _ if it is integrally closed in @xmath155 . [
prop : cluster - criterion ] let @xmath16 be a finitely generated @xmath126-algebra and a normal domain . let @xmath140 be a seed in @xmath155 satisfying the following conditions : * all elements of @xmath141 belong to @xmath16 ; * the cluster variables in @xmath141 are pairwise coprime ( in @xmath16 ) ; * for each cluster variable @xmath157 , the seed mutation @xmath133 replaces @xmath132 with an element @xmath139 ( cf . )
that lies in @xmath16 and is coprime to @xmath132 .
( here we call two elements of @xmath16 coprime if the locus of their common zeros has codimension @xmath158 in @xmath159 . )
then @xmath160 .
if , in addition , @xmath16 has a set of generators each of which appears in the seeds mutation - equivalent to @xmath140 , then @xmath161 .
proposition [ prop : cluster - criterion ] ( and its proof given below ) extends some of the ideas used in ( * ? ? ?
* proof of theorem 2.10 ) , and subsequently in ( * ? ? ? * proof of proposition 7 ) .
the argument given below actually establishes a stronger statement : under the conditions of proposition [ prop : cluster - criterion ] , the ring @xmath16 contains the _ upper cluster algebra _
@xmath162 ( see @xcite ) , or more precisely the subalgebra of @xmath155 consisting of the elements which , when expressed in terms of any extended cluster , are laurent polynomials in the cluster variables and ordinary polynomials in the coefficient variables . in particular , the conditions of proposition [ prop : cluster - criterion ] imply that in this case , the upper cluster algebra coincides with the ordinary cluster algebra ( localized at coefficient variables ) . the last implication is clear since its premise ensures that @xmath163 . to prove that @xmath160
, we need to show that each cluster variable @xmath164 from any seed @xmath165 mutation - equivalent to @xmath140 belongs to @xmath16 .
this is done using some basic algebraic geometry .
our assumptions on the ring @xmath16 mean that it can be identified with the coordinate ring of an ( irreducible ) normal affine complex algebraic variety @xmath166 .
then @xmath167 , the field of rational functions on @xmath168 .
our goal is to show that each `` distant '' cluster variable @xmath164 as above is not just a rational function in @xmath169 but a regular function in @xmath170\!=\!r$ ] .
the key property that we need is the algebraic version of hartogs continuation principle for normal varieties ( see , e.g. , ( * ? ? ?
* chapter 2 , 7.1 ) ) which asserts that a function on @xmath168 that is regular outside a closed algebraic subset of codimension @xmath158 is in fact regular everywhere on @xmath168 .
let @xmath171 be the locus where at least two of the cluster variables in @xmath141 vanish , or else one of them , say @xmath132 , vanishes together with the element @xmath139 defined by . by the coprimeness conditions imposed on the seed @xmath140 ,
the codimension of @xmath172 in @xmath168 is @xmath173 .
the complement @xmath174 consists of the points @xmath175 such that * at most one of the cluster variables in @xmath141 vanishes at @xmath78 , and * for each pair @xmath176 as above , either @xmath132 or @xmath139 does not vanish at @xmath78 .
this means that there is a seed @xmath149 ( either the original seed @xmath140 or one of the adjacent seeds @xmath177 ) none of whose cluster variables vanishes at @xmath78 ; moreover @xmath178 $ ] .
then the laurent phenomenon ( theorem [ th : laurent ] ) implies that our distant cluster variable @xmath164 is regular at @xmath78 . by the algebraic hartogs principle mentioned above
, it follows that @xmath164 is regular on @xmath168 , and we are done .
[ cor : cluster - criterion ] let @xmath16 be a finitely generated unique factorization domain over @xmath126 .
let @xmath140 be a seed in the quotient field of @xmath16 such that all elements of @xmath141 and all elements of clusters adjacent to @xmath141 are irreducible elements of @xmath16
. then @xmath160 .
if , in addition , the union of extended clusters in @xmath151 contains a generating set for @xmath16 , then @xmath161 .
let us check the conditions of proposition [ prop : cluster - criterion ] .
any ufd is a normal domain . to verify the coprimality conditions
, we first note that two elements of @xmath141 can not differ by a scalar factor since they are algebraically independent .
similarly , the elements of an exchange pair @xmath176 can not differ by a scalar factor , or else the exchange relation would give an algebraic dependence in @xmath141 .
now , an irreducible element in a ufd generates a prime ideal , so the vanishing locus in @xmath159 for each of our irreducible elements is an ( irreducible ) subvariety .
these subvarieties are pairwise distinct ( cf . above ) , so their pairwise intersections have codimensions @xmath158 . in view of lemma [ lem :
properties - of - rabv ] , in order to identify a cluster algebra structure in the ring of invariants @xmath13 , all we need to do is find a seed @xmath140 satisfying the conditions in corollary [ cor : cluster - criterion ] .
this is a lot easier said than done .
from now on , we assume that @xmath7 , so that @xmath11 .
our description of cluster structures in the classical rings of invariants @xmath13 will be based on the calculus of tensor diagrams @xcite , particularly in its version of the `` @xmath179 spider '' developed by g. kuperberg @xcite .
this calculus , in turn , is based on two simple observations . the first one was already mentioned in connection with the first fundamental theorem : * the identity tensor and the volume tensors on @xmath1 and @xmath21 are @xmath0-invariant .
the second observation is that * the operations of tensor product and contraction preserve @xmath0 invariance . by combining the ingredients derived from ( i ) using the operations in ( ii )
, one can build a surprisingly rich supply of @xmath0-invariant tensors and from them , a large class of polynomial invariants .
we next explain how , without claiming any originality : the machinery described below is a variation of diagrammatic calculus whose various versions were ( re)discovered multiple times , under different guises and names , see in particular @xcite .
while tensors themselves are multilinear functionals , they can be used to construct invariants of higher degrees using _ cloning _ ( or _ restitution _ ) , the procedure opposite to _
polarization_. the idea is extremely simple : evaluate an @xmath0-invariant tensor at a tuple of arguments ( vectors and/or covectors ) some of which are repeated several times .
nontrivial invariants constructed using this technique will appear in ...
tensor diagrams are built using three types of building blocks shown in figure [ fig : webs21 ] .
these blocks correspond , left to right , to the three basic @xmath0-invariant tensors : the volume tensor , the dual volume tensor , and the identity tensor .
the endpoints of a block correspond to the tensor s arguments : a sink to a vector , a source to a covector . in order for either of the volume tensors to be well defined , a cyclic ordering of the three arguments must be specified .
( 20,20)(0,0 ) ( 10,10)(1,1)10 ( 10,10)(-1,1)10 ( 10,10)(0,-1)10 ( 20,20)(0,0 ) ( 0,20)(1,-1)10 ( 20,20)(-1,-1)10 ( 10,0)(0,1)10 ( 15,20)(0,0 ) ( 0,10)(1,0)15 we then combine these building blocks by plugging arrowheads into arrowtails . interpreting this operation as contraction of the corresponding tensors ,
we obtain a well defined notion of a tensor associated with the resulting combinatorial gadget . note that plugging one end of a two - pronged block ( corresponding to the identity tensor ) into an end of opposite valence in some tensor diagram does not change the tensor that the latter represents .
consequently , for any diagram assembled using the above procedure from the three basic types of blocks , the associated tensor does not depend on the locations ( or number ) of plug - in joints along unbranched stretches . erasing those joints produces a directed graph with trivalent and univalent vertices in which every trivalent vertex is a source or a sink .
any such graph , together with the additional data specifying cyclic ordering of edges incident to trivalent vertices , represents a well defined tensor .
we are going to draw all our diagrams in an oriented disk , obeying the following conventions .
we will usually erase the arrows , introducing instead a bi - coloring of the vertices : the sinks will be colored black , and the sources white .
we place the boundary vertices on the boundary of the disk , in the order mirroring the one used to define the direct product of the @xmath104 spaces @xmath1 and @xmath21 whose @xmath0 invariants we study .
we place the interior vertices inside the disk , and draw the edges as simple curves that are allowed to intersect transversally .
last but not least , we make sure that the cyclic ordering specified at each interior vertex of the diagram matches the clockwise ordering of the three curves meeting at the corresponding point in the disk .
the last step in the construction of a general tensor diagram involves gluing together some of the univalent vertices of a directed graph as above .
we interpret this step as substituting the same ( co)vector into the arguments of the associated tensor which correspond to the glued endpoints . at this point ,
the @xmath0 invariant represented by the graph ceases to be multilinear .
this last operation is sometimes called ( partial ) _ restitution _ , and is inverse to _
polarization_. an example illustrating different stages of the construction outlined above is shown in figure [ fig : webs22 ] .
( 60,67)(0,-7 ) ( 10,30)(-1,0)10 ( 10,30)(1,1)9 ( 10,30)(2,-3)8.5 ( 20,15)(2,-3)8 ( 50,30)(1,0)10 ( 50,30)(-1,1)9 ( 50,30)(-2,-3)8.5 ( 40,15)(-2,-3)8 ( 21,41)(1,1)9 ( 39,41)(-1,1)9 ( 30,60)(0,-1)10 ( 26.5,-7)(a ) ( 60,67)(0,-7 ) ( 10,30)(-1,0)10 ( 10,30)(1,1)20 ( 10,30)(2,-3)18 ( 50,30)(1,0)10 ( 50,30)(-1,1)20 ( 50,30)(-2,-3)18 ( 30,60)(0,-1)10 ( 26.5,-7)(b ) + ( 60,67)(0,-7 ) ( 0,30)(1,0)8.5 ( 30,50)(-1,-1)19 ( 30,50)(1,-1)19 ( 28,3)(-2,3)17.3 ( 60,30)(-1,0)9 ( 32,3)(2,3)17.3 ( 30,59)(0,-1)9 ( 0,30 ) ( 60,30 ) ( 30,50 ) ( 28,3 ) ( 32,3 ) ( 10,30 ) ( 50,30 ) ( 30,60 ) ( 26.5,-8.5)(c ) ( 60,67)(0,-7 ) ( 0,30)(1,0)8.5 ( 30,50)(-1,-1)19 ( 30,50)(1,-1)19 ( 30,0)(-2,3)19.3 ( 60,30)(-1,0)9 ( 30,0)(2,3)19.3 ( 30,59)(0,-1)9 ( 0,30 ) ( 60,30 ) ( 30,50 ) ( 30,0 ) ( 10,30 ) ( 50,30 ) ( 30,60 ) ( 45,10)@xmath180 ( 26.5,-8.8)(d ) a _ tensor diagram _ is a finite bipartite graph @xmath180 with a fixed proper coloring of its vertices into two colors , black and white , and with a fixed partition of its vertex set into the set @xmath181 of _ boundary _ vertices and the set @xmath182 of _ internal _ vertices , satisfying the following conditions : * each internal vertex is trivalent ; * for each internal vertex , a cyclic order on the edges incident to it is fixed .
the last condition ensures that each of the six permutations of the three edges incident to an internal vertex has a well - defined sign .
if @xmath181 consists of @xmath15 white vertices and @xmath14 black ones , then we say that @xmath180 is of _ type _ @xmath23 .
we regard each edge of a tensor diagram as being oriented so that it points away from its white endpoint and towards the black one . while drawing tensor diagrams , this orientation is routinely omitted .
a tensor diagram @xmath180 of type @xmath23 defines an @xmath0 invariant @xmath183\in r_{a , b}(v)$ ] obtained by repeated contraction of elementary @xmath0-invariant tensors ( followed by bundling up some of the arguments ) , as explained above . to compensate for the informality of that explanation
, we next provide a precise definition of the invariant @xmath183 $ ] in the language of coordinates .
let us identify @xmath13 with the ring of @xmath184 invariants of collections of @xmath15 covectors @xmath185 and @xmath14 vectors @xmath186 labeled by a fixed collection of @xmath15 white and @xmath14 black vertices on the boundary of a disk .
the invariant @xmath183 $ ] associated with a tensor diagram @xmath180 with those boundary vertices is then given by @xmath187-in - coordinates } [ d]=\sum_\ell \biggl(\,\prod_{v\in\operatorname{int}(d)}\operatorname{sign}(\ell(v))\biggr ) \biggl(\,\prod_{\substack{v\in\operatorname{bd}(d)\\
\text{$v$ black}}}x(v)^{\ell(v)}\biggr ) \biggl(\,\prod_{\substack{v\in\operatorname{bd}(d)\\ \text{$v$ white}}}y(v)^{\ell(v)}\biggr),\ ] ] where * @xmath188 runs over all proper labelings of the edges in @xmath180 by the numbers @xmath189 ; to wit , we require that for each internal vertex @xmath53 , the labels associated with the three edges incident to @xmath53 are all distinct ; * @xmath190 denotes the sign of the ( cyclic ) permutation of those three labels determined by the cyclic ordering of the edges incident to @xmath53 ; * @xmath191 denotes the monomial @xmath192 , product over all edges @xmath193 incident to @xmath53 , and similarly for @xmath194 .
example [ example : tripod - and - stick ] below illustrates formula . many more examples of tensor diagrams are scattered throughout the paper .
[ example : tripod - and - stick ] consider the tensor diagram @xmath180 shown in figure [ fig : webs22](d ) .
we label the black boundary vertices 1 , 2 , and 3 , in clockwise order , and label the white vertex 4 .
we denote the corresponding three column vectors and one row covector by @xmath195 respectively .
there are 24 proper labelings of the edges of @xmath180 .
they can be broken into four six - tuples as shown in figure [ fig : edge - labelings ] .
the corresponding monomial weights are @xmath196 where @xmath197 denotes the sign of the permutation @xmath198 in the symmetric group @xmath199 .
the weights of the labelings in the fourth group cancel each other out under the sign - reversing involution that switches @xmath200 and @xmath201 .
consequently , becomes @xmath202&=\sum_{(\alpha,\beta,\gamma)\in\mathcal{s}_3 } \operatorname{sgn}(\alpha,\beta,\gamma ) x_{\gamma 1}x_{\alpha 2}x_{\beta 3 } \,(x_{\alpha 2 } y_\alpha+x_{\beta 2 } y_\beta+x_{\gamma 2 } y_\gamma)\\ \label{eq : det*dot - product } & = \det(x_{ij})_{i , j\in(1,2,3 ) } ( x_{12 } y_1+x_{22 } y_2+x_{32 } y_3 ) , \end{aligned}\ ] ] coinciding with the invariant given by the tensor diagram shown in figure [ fig : tripod - and - a - stick ] .
( 60,60)(0,0 ) ( 0,30)(1,0)8.5 ( 30,50)(-1,-1)19 ( 30,50)(1,-1)19 ( 30,0)(-2,3)19.3 ( 60,30)(-1,0)9 ( 30,0)(2,3)19.3 ( 30,59)(0,-1)9 ( 0,30 ) ( 60,30 ) ( 30,50 ) ( 30,0 ) ( 10,30 ) ( 50,30 ) ( 30,60 ) ( 0,25)(0,0)@xmath2 ( 60,25)(0,0)@xmath203 ( 25,0)(0,0)@xmath173 ( 25,60)(0,0)@xmath204 ( 15,15)(0,0)@xmath200 ( 45,15)(0,0)@xmath200 ( 5,35)(0,0)@xmath205 ( 55,35)(0,0)@xmath201 ( 18,43)(0,0)@xmath201 ( 42,43)(0,0)@xmath205 ( 34,55)(0,0)@xmath200 ( 60,60)(0,0 ) ( 0,30)(1,0)8.5 ( 30,50)(-1,-1)19 ( 30,50)(1,-1)19 ( 30,0)(-2,3)19.3 ( 60,30)(-1,0)9 ( 30,0)(2,3)19.3 ( 30,59)(0,-1)9 ( 0,30 ) ( 60,30 ) ( 30,50 ) ( 30,0 ) ( 10,30 ) ( 50,30 ) ( 30,60 ) ( 0,25)(0,0)@xmath2 ( 60,25)(0,0)@xmath203 ( 25,0)(0,0)@xmath173 ( 25,60)(0,0)@xmath204 ( 15,15)(0,0)@xmath200 ( 45,15)(0,0)@xmath205 ( 5,35)(0,0)@xmath205 ( 55,35)(0,0)@xmath201 ( 18,43)(0,0)@xmath201 ( 42,43)(0,0)@xmath200 ( 34,55)(0,0)@xmath205 ( 60,60)(0,0 ) ( 0,30)(1,0)8.5 ( 30,50)(-1,-1)19 ( 30,50)(1,-1)19 ( 30,0)(-2,3)19.3 ( 60,30)(-1,0)9 ( 30,0)(2,3)19.3 ( 30,59)(0,-1)9 ( 0,30 ) ( 60,30 ) ( 30,50 ) ( 30,0 ) ( 10,30 ) ( 50,30 ) ( 30,60 ) ( 0,25)(0,0)@xmath2 ( 60,25)(0,0)@xmath203 ( 25,0)(0,0)@xmath173 ( 25,60)(0,0)@xmath204 ( 15,15)(0,0)@xmath201 ( 45,15)(0,0)@xmath200 ( 5,35)(0,0)@xmath205 ( 55,35)(0,0)@xmath201 ( 18,43)(0,0)@xmath200 ( 42,43)(0,0)@xmath205 ( 34,55)(0,0)@xmath201 ( 60,60)(0,0 ) ( 0,30)(1,0)8.5 ( 30,50)(-1,-1)19 ( 30,50)(1,-1)19 ( 30,0)(-2,3)19.3 ( 60,30)(-1,0)9 ( 30,0)(2,3)19.3 ( 30,59)(0,-1)9 ( 0,30 ) ( 60,30 ) ( 30,50 ) ( 30,0 ) ( 10,30 ) ( 50,30 ) ( 30,60 ) ( 0,25)(0,0)@xmath2 ( 60,25)(0,0)@xmath203 ( 25,0)(0,0)@xmath173 ( 25,60)(0,0)@xmath204 ( 15,15)(0,0)@xmath200 ( 45,15)(0,0)@xmath201 ( 5,35)(0,0)@xmath205 ( 55,35)(0,0)@xmath205 ( 18,43)(0,0)@xmath201 ( 42,43)(0,0)@xmath200 ( 34,55)(0,0)@xmath205 ( 60,60)(0,0 ) ( 0,30)(1,0)18.5 ( 30,0)(-1,3)9.5 ( 60,30)(-1,0)39 ( 30,59)(0,-1)59 ( 0,30 ) ( 60,30 ) ( 30,0 ) ( 20,30 ) ( 30,60 ) ( 0,25)(0,0)@xmath2 ( 60,25)(0,0)@xmath203 ( 25,0)(0,0)@xmath173 ( 25,58)(0,0)@xmath204 the @xmath0 invariant @xmath183 $ ] defined by a tensor diagram @xmath180 is homogeneous in each of its arguments @xmath206 , for @xmath53 a boundary vertex
. the degree of @xmath183 $ ] with respect to @xmath206 is equal to the degree @xmath207 of @xmath53 ( in the graph theory sense ) .
the _ multidegree _ of @xmath180 is , by definition , the multidegree of @xmath183 $ ] ( cf . ):
@xmath208)= ( d_v)_{v\in\operatorname{bd}(d)}.\ ] ] an @xmath10 invariant @xmath183 $ ] associated with a tensor diagram @xmath180 without multiple edges can be alternatively defined by a determinantal formula , as explained below .
although we do not rely on such formulas in this paper , they may be useful in some calculations .
as before , suppose that @xmath180 has type @xmath23 .
assume that no two boundary vertices are directly connected by an edge .
( otherwise , factor out the corresponding scalar product , and remove the edge . ) assume that @xmath180 has @xmath209 interior vertices , @xmath210 of them white and @xmath211 black .
clearly @xmath212 , the number of edges in @xmath180 .
we now build a square @xmath213 matrix @xmath214 as follows : * the rows of @xmath214 come from the white vertices in @xmath182 ( each such vertex contributes three rows ) as well as from the edges incident to the white vertices in @xmath181 ( each such edge contributes one row ) ; * similarly , the columns of @xmath214 come from the black interior vertices ( each contributes three ) and from the edges incident to the black boundary vertices ( each contributes one ) .
the matrix @xmath214 is then built out of rectangular blocks @xmath215 of sizes @xmath216 , @xmath217 , @xmath218 , and @xmath219 , defined as follows : * if @xmath87 is a white interior vertex and @xmath88 is a black interior vertex , then @xmath215 is a @xmath216 identity matrix ; * if @xmath87 is a white interior vertex and @xmath88 is an edge incident to a black boundary vertex @xmath53 , then @xmath215 is the column vector @xmath206 ; * if @xmath88 is a black interior vertex and @xmath87 is an edge incident to a white boundary vertex @xmath53 , then @xmath215 is the row ( co)vector @xmath220 ; * otherwise , @xmath215 consists entirely of zeroes .
it is straightforward to check that @xmath221 $ ] .
( the sign depends on the ordering of the rows and columns in @xmath214 . )
it is also easy to see directly that @xmath222 is an @xmath223 invariant .
we illustrate this construction for the tensor diagram shown in figure [ fig : webs22](d ) , using the notation introduced in example [ example : tripod - and - stick ] .
the recipe described above gives @xmath224.\ ] ] it is then easy to see that @xmath225 matching .
the calculus of tensor diagrams includes natural counterparts for both the additive and the multiplicative structures in the ring of invariants .
for the former , allow formal linear combinations of tensor diagrams , and extend the definition of @xmath183 $ ] by linearity .
for the latter , use superposition of diagrams : if @xmath180 is a union of subdiagrams @xmath226 connected only at boundary vertices , then @xmath183=[d_1][d_2]\cdots$ ] ; cf .
figure [ fig : tripod - and - a - stick ] .
the weyl generators of the ring @xmath13 are encoded by the simplest possible tensor diagrams : two types of _ tripods _ ( corresponding to two kinds of plcker coordinates ) and simple edges .
to rephrase , one takes each of the three types of building blocks shown in figure [ fig : webs21 ] and attaches it directly to the boundary vertices .
the first fundamental theorem implies that any @xmath0 invariant can be represented ( non - uniquely ) as a linear combination of invariants associated with tensor diagrams obtained by superposition of tripods and edges . an important class of relations among invariants associated with tensor diagrams on an oriented plane is obtained from _ local transformation rules_. such a rule transforms a small fragment @xmath227 of a tensor diagram @xmath180 into a linear combination @xmath228 of other such pieces ( corresponding to tensors of the same type ) while keeping the rest of the diagram intact .
the intrinsic definition of the invariants @xmath183 $ ] implies that if @xmath229\!=\!\sum_i c_i [ f_i]$ ] ( i.e. , the tensor does not change locally ) , then @xmath183=\sum_i c_i [ d_i]$ ] , where @xmath230 denotes the tensor diagram obtained from @xmath180 by replacing @xmath227 by @xmath231 . figure [ fig : skein ] shows several fundamental relations of this kind . checking their validity is straightforward .
several basic relations involving boundary vertices are shown in figure [ fig : local - boundary ] .
cl ( 16,20)(0,0 ) ( 8,10)(0,0)(a ) & ( 30,20)(0,0 ) ( 0,20)(3,-2)30 ( 0,20)(3,-2)25 ( 0,0)(3,2)30 ( 0,0)(3,2)25 ( 16,20)(0,0 ) ( 8,10)(0,0)@xmath232 ( 30,20)(0,0 ) ( 0,0)(1,1)9.2 ( 0,20)(1,-1)9.2 ( 10,10 ) ( 11,10)(1,0)8 ( 20,10 ) ( 30,20)(-1,-1)9.2 ( 30,0)(-1,1)9.2 ( 16,20)(0,0 ) ( 8,10)(0,0)@xmath233 ( 30,20)(0,0 ) ( 0,0)(1,0)30 ( 0,0)(1,0)20 ( 0,20)(1,0)30 ( 0,20)(1,0)20 + ( 16,30)(0,0 ) ( 8,15)(0,0)(b ) & ( 30,30)(0,0 ) ( 0,0)(1,1)9 ( 30,30)(-1,-1)9 ( 0,30)(1,-1)9 ( 30,0)(-1,1)9 ( 10,10 ) ( 20,20 ) ( 10,20 ) ( 20,10 ) ( 11,10)(1,0)8 ( 11,20)(1,0)7.8 ( 10,11.2)(0,1)8 ( 20,11)(0,1)8 ( 16,30)(0,0 ) ( 8,15)(0,0)@xmath232 ( 30,30)(0,0 ) ( 0,0)(1,0)30 ( 15,0)(-1,0)5 ( 0,30)(1,0)30 ( 15,30)(1,0)5 ( 16,30)(0,0 ) ( 8,15)(0,0)@xmath233 ( 30,30)(0,0 ) ( 0,0)(0,1)30 ( 0,15)(0,-1)5 ( 30,0)(0,1)30 ( 30,15)(0,1)5 + ( 16,10)(0,0 ) ( 8,5)(0,0)(c ) & ( 30,10)(0,0 ) ( 0,5)(1,0)9 ( 30,5)(-1,0)9 ( 11,6)(15,10)(19,6 ) ( 11,4)(15,0)(19,4 ) ( 10,5 ) ( 20,5 ) ( 16,10)(0,0 ) ( 8,5)(0,0)@xmath232 ( 14,10)(0,0 ) ( 6,5)(0,0)@xmath234 ( 30,10)(0,0 ) ( 0,5)(1,0)30 ( 15,5)(1,0)5 + ( 16,20)(0,0 ) ( 8,10)(0,0)(d ) & ( 30,20)(0,0 ) ( 15,10 ) ( 16,20)(0,0 ) ( 8,10)(0,0)@xmath232 ( 4,20)(0,0 ) ( 1.5,10.5)(0,0)@xmath2 cl ( 16,25)(0,-5 ) ( 8,10)(0,0)(a ) & ( 10,35)(0,-5 ) ( 5,-5)(0,0)@xmath235 ( 5,0 ) ( 5,20 ) ( 5,21)(0,1)8 ( 4,1)(-3,10)(4,19 ) ( 6,1)(13,10)(6,19 ) ( 12,25)(0,-5 ) ( 6,10)(0,0)@xmath232 ( 10,25)(0,-5 ) ( 5,-5)(0,0)@xmath235 ( 5,0 ) ( 5,20 ) ( 5,21)(0,1)8 ( 4,1)(-3,10)(4,19 ) ( 6,1)(13,10)(6,19 ) ( 16,25)(0,-5 ) ( 6,10)(0,0)@xmath236 + ( 16,25)(0,-5 ) ( 8,10)(0,0)(b ) & ( 10,25)(0,-5 ) ( 5,-5)(0,0)@xmath235 ( 5,0 ) ( 5,20 ) ( 5,1)(0,1)18 ( 4,1)(-3,10)(4,19 ) ( 6,1)(13,10)(6,19 ) ( 12,25)(0,-5 ) ( 6,10)(0,0)@xmath232 ( 10,25)(0,-5 ) ( 5,-5)(0,0)@xmath235 ( 5,0 ) ( 5,20 ) ( 5,1)(0,1)18 ( 4,1)(-3,10)(4,19 ) ( 6,1)(13,10)(6,19 ) ( 16,25)(0,-5 ) ( 6,10)(0,0)@xmath236 figure [ fig : yang - baxter ] shows a few additional relations which are immediate from the definition of the invariants associated with tensor diagrams .
these relations can also be deduced from those in figure [ fig : skein ] .
cl ( 16,20)(0,0 ) ( 8,10)(0,0)(a ) & ( 30,20)(0,0 ) ( 0,20)(3,-2)30 ( 0,0)(3,2)30 ( 0,10)(15,0)(30,10 ) ( 16,20)(0,0 ) ( 8,10)(0,0)@xmath232 ( 30,20)(0,0 ) ( 0,20)(3,-2)30 ( 0,0)(3,2)30 ( 0,10)(15,20)(30,10 ) + ( 16,16)(0,0 ) ( 8,8)(0,0)(b ) & ( 30,16)(0,0 ) ( 0,0)(20,5)(20,10 ) ( 30,0)(10,5)(10,10 ) ( 10,10)(10,15)(15,15 ) ( 20,10)(20,15)(15,15 ) ( 16,16)(0,0 ) ( 8,8)(0,0)@xmath232 ( 30,16)(0,0 ) ( 0,0)(1,0)30 + ( 16,10)(0,0 ) ( 8,5)(0,0)(c ) & ( 30,10)(0,0 ) ( 0,0)(15,15)(30,0 ) ( 0,10)(15,-5)(30,10 ) ( 16,10)(0,0 ) ( 8,5)(0,0)@xmath232 ( 30,10)(0,0 ) ( 0,0)(1,0)30 ( 0,10)(1,0)30 + ( 16,20)(0,0 ) ( 8,10)(0,0)(d ) & ( 30,20)(0,0 ) ( 0,10)(1,0)8.8 ( 10,10 ) ( 11,11)(20,18)(30,0 ) ( 11,9)(20,2)(30,20 ) ( 16,20)(0,0 ) ( 8,10)(0,0)@xmath232 ( 12,20)(0,0 ) ( 5,10)(0,0)@xmath237 ( 30,20)(0,0 ) ( 0,10)(1,0)8.8 ( 10,10 ) ( 30,20)(-2,-1)18.7 ( 30,0)(-2,1)18.7 + ( 16,20)(0,0 ) ( 8,10)(0,0)(e ) & ( 30,20)(0,0 ) ( 0,10)(1,0)8.8 ( 10,10 ) ( 11,11)(20,18)(30,0 ) ( 11,9)(20,2)(30,20 ) ( 16,20)(0,0 ) ( 8,10)(0,0)@xmath232 ( 12,20)(0,0 ) ( 5,10)(0,0)@xmath237 ( 30,20)(0,0 ) ( 0,10)(1,0)8.8 ( 10,10 ) ( 30,20)(-2,-1)18.7 ( 30,0)(-2,1)18.7 it is possible to develop a diagrammatic calculus based on the local relations listed above , without making any mention of tensors or the special linear group @xmath10 ; see in particular @xcite and references therein .
such diagrammatic calculus , and its @xmath238-analogues , can in particular be used to construct invariants of knots and links .
we do not discuss these connections in this paper .
informally speaking , webs are _ planar _ tensor diagrams .
the systematic study of webs was pioneered by g. kuperberg @xcite whose foundational results are reviewed below , in slightly different terminology .
a ( planar ) _ web _ is a tensor diagram @xmath180 embedded in an oriented disk as described above in section [ sec : tensor - diagrams ] , so that its edges do not cross or touch each other , except at endpoints . each web is considered up to an isotopy of the disk that fixes its boundary so it is in essence a combinatorial object .
recall that the boundary ( respectively internal ) vertices of @xmath180 must lie on the boundary ( respectively in the interior ) of the disk , and the three edges meeting at each internal vertex are viewed as cyclically ordered clockwise .
a web is called _ non - elliptic _ if it has no multiple edges , and no unoriented simple 4-cycles whose all four vertices are internal . an invariant @xmath183 $ ] associated with a non - elliptic web @xmath180 is called a _ web invariant_. a web @xmath180 of type @xmath23 has @xmath15 white boundary vertices and @xmath14 black ones . the cyclic pattern of colors of the boundary vertices is encoded by the ( cyclic ) _ signature _ of @xmath180 ( cf . a similar notion introduced in section [ sec : rings - of - invariants ] ) , a cyclically ordered binary string , or more precisely a word in the alphabet @xmath103 considered up to a cyclic permutation .
for example , the webs in figures [ fig : edge - labelings][fig : tripod - and - a - stick ] have signature @xmath239 = [ \bullet\bullet\circ\,\bullet ] = [ \bullet\circ\bullet\,\bullet ] = [ \circ\bullet\bullet\,\bullet]\ ] ] the following highly nontrivial result is the cornerstone of the theory of webs .
[ th : web - basis ] web invariants with a fixed signature @xmath112 of type @xmath23 form a linear basis in the ring of invariants @xmath113 .
being a special kind of tensor diagrams , webs have _ multidegrees _
, cf . ; the invariants they define are multi - homogeneous .
kuperberg s theorem can be restated as saying that web invariants with a fixed signature @xmath112 and a fixed multidegree make up a linear basis in the corresponding multi - homogeneous component of @xmath111 .
[ example : r24 ] for the ring of invariants @xmath240 , choose the signature @xmath241\ ] ] and the multidegree @xmath242 .
there are @xmath243 non - elliptic webs of such signature and multidegree , shown in figure [ fig : webs23 ] at the top .
thus the corresponding multi - homogeneous component of @xmath244 is @xmath243-dimensional .
theorem [ th : web - basis ] implies that the number of non - elliptic webs of given multidegree does not depend on the choice of a signature of given type @xmath23 . to illustrate ,
continue with example [ example : r24 ] .
taking instead the signature @xmath245 $ ] of the same type @xmath246 , and choosing the multidegree @xmath247so that , as before , the invariants in question are multilinear in all arguments except for a single covector we get the @xmath243 webs ( same number ! ) shown in figure [ fig : webs23 ] at the bottom .
[ rem : confluence ] any linear combination of tensor diagrams can be transformed into a linear combination of non - elliptic webs by repeated application of local relations shown in figures .
( just apply them left - to - right . )
it follows that @xmath111 is spanned by the web invariants .
the content of theorem [ th : web - basis ] is that web invariants are linearly independent , so the reduction process described above is _
confluent_. theorem [ th : web - basis ] implies that two linear combinations of tensor diagrams define the same invariant if and only if they can be transformed into each other using the relations in figures [ fig : skein][fig : local - boundary ] .
[ rem : khovanov - kuperberg ] it is tempting to hypothesize , as m. khovanov and g. kuperberg originally did @xcite , that the web basis in a space of multilinear invariants ( of some fixed signature ) coincides with the corresponding instance of g. lusztig s _ dual canonical basis _ ( see @xcite for definitions and references ) .
their investigation however established that web bases for @xmath10 are generally _ not _ dual canonical .
the first discrepancy occurs in degree 12 , for the signature @xmath248 .
see section [ sec : thickening - non - cluster ] for a cluster - theoretic interpretation of this counterexample .
a _ closed _ tensor diagram ( without boundary vertices ) represents a tensor of type @xmath249 , _ i.e. _ a scalar .
this scalar has the following direct description .
[ prop : closed - webs ] let @xmath180 be a closed web with @xmath250 white and @xmath250 black vertices .
then @xmath183 $ ] is equal to @xmath251 times the number of proper colorings of the edges of @xmath180 into three colors .
( that is , we require the colors of the three edges incident to each vertex of @xmath180 to be distinct . ) the simplest illustration of proposition [ prop : closed - webs ] is provided by formula ( d ) in figure [ fig : skein ] ( with @xmath252 ) . a more interesting example
: let @xmath180 be the @xmath203-skeleton of the ( three - dimensional ) cube ; then @xmath183=24 $ ] .
the proof of proposition [ prop : closed - webs ] is omitted ; we do not rely on it elsewhere in the paper . a few more are shown in figure [ fig : closed - webs ] .
ccccc ( 10,20)(0,0 ) ( 5,0 ) ( 5,20 ) ( 5,1)(0,1)18 ( 4,1)(-3,10)(4,19 ) ( 6,1)(13,10)(6,19 ) & & ( 30,20)(0,0 ) ( 5,0 ) ( 5,20 ) ( 6.5,20)(1,0)17.5 ( 4,1)(0,10)(4,19 ) ( 6,1)(10,10)(6,19 ) ( 25,0 ) ( 25,20 ) ( 6,0)(1,0)18 ( 24,1)(20,10)(24,19 ) ( 26,1)(30,10)(26,19 ) & & ( 20,20)(0,0 ) ( 6,6 ) ( 14,14 ) ( 6,14 ) ( 14,6 ) ( 0,0 ) ( 20,20 ) ( 0,20 ) ( 20,0 ) ( 1,0)(1,0)17.5 ( 1,20)(1,0)18 ( 0,1)(0,1)18 ( 20,1.5)(0,1)17.5 ( 7,6)(1,0)6 ( 7,14)(1,0)5.5 ( 6,7.5)(0,1)6 ( 14,7)(0,1)6 ( 5,5)(-1,-1)5 ( 15,15)(1,1)5 ( 6,14)(-1,1)5 ( 14,6)(1,-1)5 + @xmath183=-6 $ ] & & @xmath183=12 $ ] & & @xmath183=24 $ ]
fix a ( cyclic ) signature @xmath112 of type @xmath23 with @xmath253 .
we assume that @xmath112 is _ non - alternating _ , i.e. , it has two adjacent vertices of the same color . in this section ,
we construct a family of `` special '' elements in the ring of invariants @xmath113 .
these special invariants will play a key role in our main construction .
the proofs of the propositions stated in sections [ sec : special - invariants][sec : special - seeds ] are given in section [ sec : special - proofs ] .
as before , we work in a disk with @xmath104 marked points ( vertices ) on the boundary , @xmath15 of them white and @xmath14 of them black , arranged in accordance with the signature @xmath112 .
we label these vertices @xmath254 , going clockwise , and work with them modulo @xmath104 . for each boundary vertex @xmath255 , we next define two trees @xmath256 and @xmath257 embedded into our disk .
these trees will then serve as building blocks for certain tensor diagrams . if the vertex @xmath255 is black , then @xmath256 has one vertex , namely @xmath255 , and no edges .
if @xmath255 is white , then place a new black vertex ( which we call the _ proxy _ of @xmath255 ) inside the disk and connect it to @xmath255 .
we then examine @xmath258 .
if @xmath258 is white , then we connect it to the proxy vertex , and stop ; see figure [ fig : caterpillars](a )
. if @xmath255 is white and @xmath259 is black , then we look at @xmath260 .
if @xmath260 is black , then put a white vertex inside the disk , connect it to both @xmath258 and @xmath260 , and to the proxy vertex ; see figure [ fig : caterpillars](b ) .
if @xmath255 is white , @xmath258 is black , and @xmath260 is white , then examine @xmath261 , and so on , cf . figure [ fig : caterpillars](c ) . in general ,
we proceed clockwise from @xmath255 until we find two consecutive vertices of the same color ( here we need the condition that @xmath112 is non - alternating ) , then build a caterpillar - like bi - colored tree whose all interior vertices have degree @xmath2 except for the proxy vertex which has degree @xmath173 . the graph @xmath257 is defined in the same way , with the colors swapped .
the proxy vertex is defined analogously ( if @xmath255 is white ) . at the next stage , we stitch together several caterpillar trees to build tensor diagrams . in anticipation of this stage , we draw these trees without self - intersections , and sufficiently close to the boundary of the disk , so as to avoid ambiguities later on related to the choice of cyclic ordering at each vertex of the tensor diagram .
[ def : special - inv ] let @xmath255 and @xmath238 be boundary vertices , @xmath262 .
the special invariant @xmath263 is defined by the tensor diagram obtained by connecting the trees @xmath256 and @xmath264 by a single edge .
one of its endpoints is @xmath255 if the latter is black , or else take the proxy of @xmath255 .
the other endpoint is @xmath238 if the latter is white , or else take the proxy of @xmath238 .
make sure the connector edge approaches every proxy vertex from the same side where the disk s center lies .
see figure [ fig : webs61 ] for a couple of examples .
now , let @xmath265 be distinct boundary vertices , ordered clockwise .
the special invariants @xmath266 and @xmath267 are defined by tensor diagrams constructed in a similar fashion to those used for @xmath263 . for @xmath266 , place a white vertex in the middle , and draw edges from it to each of @xmath268 . as the other endpoints of these three edges , use the vertices @xmath265 whenever they are black , or else take their respective proxies . for @xmath267 ,
reverse the roles of the colors .
see figure [ fig : webs61 ] .
finally , let @xmath269 be four boundary vertices , ordered clockwise .
the invariant @xmath270 is defined by the tensor diagram obtained as follows .
place a white vertex @xmath271 and a black vertex @xmath235 near the center of the disk .
connect them by an edge .
connect @xmath271 to @xmath272 and @xmath273 using @xmath255 and @xmath238 if these two are black , or else using their proxies as needed .
similarly , connect @xmath235 to the appropriate white vertices in @xmath274 and @xmath275 .
make sure that the five edges incident to @xmath235 and @xmath271 do not cross each other
. + recall that @xmath1 is a three - dimensional vector space endowed with a volume form .
cross product _ of two vectors @xmath276 is the covector @xmath277 given by @xmath278 , for @xmath279 .
one similarly defines a cross product of two covectors in @xmath21 ( which is a vector ) .
[ def : partner ] consider a collection of @xmath15 covectors and @xmath14 vectors arranged around a circle in the order specified by the non - alternating signature @xmath112 .
we use the following recursive procedure to construct @xmath15 vectors and @xmath14 covectors whom we call _ partners _ of their respective counterparts .
the partner @xmath280 of a vector @xmath53 is given by @xmath281 where @xmath282 is the clockwise neighbor of @xmath53 if this neighbor is a vector ; otherwise @xmath282 is the neighbor s partner .
similarly , for each covector take its cross product with the clockwise neighbor or with the neighbor s partner .
this rule uniquely determines a partner for each vector and covector . to find all the partners , locate two consecutive vectors or two consecutive covectors ( these exist since @xmath112 is non - alternating . ) starting there , proceed in counterclockwise order .
[ example : ooxxox ] let @xmath283 $ ] , with covectors @xmath284 and vectors @xmath285 arranged in a circle as shown in figure [ fig : webs51 ] .
their partners are determined as follows : @xmath286 the left part of figure [ fig : webs51 ] illustrates the recursive process of computing the partners .
note that this kind of picture is _ not _ a tensor diagram ( although its portions can and will be used to construct web invariants ) .
rather , it can be viewed as an _ incidence diagram _ whose black and white vertices represent vectors and covectors , respectively , and whose edges connect a vector @xmath53 with a covector @xmath54 whenever @xmath287 .
projectivizing , we get a configuration in the projective plane consisting of points and lines corresponding to vectors and covectors , respectively . in an incidence diagram , a black vertex representing a point is connected to a white vertex representing a line whenever the point lies on the line .
each unordered pair @xmath288 consists of a vector and a covector annihilating each other .
projectively , they represent a _ flag _ in the plane .
furthermore , our definition of partners imposes certain incidence relations across these flags , whereby a line from one flag is forced to contain a point from another flag .
it is not difficult to see that these relations encode configurations which can be called _ necklaces of flags _ ; such a necklace is a cyclically ordered collection of flags in which each pair of consecutive flags is connected by one incidence relation ( the point in one flag lies on the line in another flag ) .
the specific form of these relations is determined by the signature @xmath112 .
see figure [ fig : webs51 ] .
the _ special invariants _ are tensor diagram invariants @xmath289 and @xmath267 defined as follows : * for @xmath290 , the invariant @xmath263 is obtained by pairing the vector in @xmath291 with the covector in @xmath292 . * for @xmath293 ,
the invariant @xmath266 ( respectively @xmath267 ) is defined as the signed volume of the three vectors ( respectively the three covectors ) in the 6-tuple @xmath294 . by convention , @xmath295 , and similarly for the upper indices . in example
[ example : ooxxox ] , we have @xmath296 some special invariants are identically equal to zero . [ prop : special=0 ] if the vertices @xmath255 and @xmath238 are not adjacent , then @xmath263 is a nontrivial invariant . if @xmath255 and @xmath238 are adjacent , then exactly one of the two invariants @xmath263 and @xmath297 vanishes .
specifically , if @xmath255 is white , then @xmath298 ; if @xmath255 is black , then @xmath299 . for example ,
in figure [ fig : webs61 ] we have @xmath300 .
some nonzero special invariants are equal to each other .
we do not catalogue all such instances ( although we could ) .
the caption to figure [ fig : webs61 ] .
[ prop : specweb ] each nonzero special invariant is a web invariant . to illustrate , refer to figure [ fig : webs61 ] . while the two tensor diagrams in the top row are non - elliptic webs , the ones in the bottom row are not .
they can however be transformed using skein relations into the non - elliptic webs shown in figure [ fig : webs62 ] .
we call a nonzero special invariant _ indecomposable _ if it does not factor as a product of two or more special invariants . in our running example of signature , the invariants
@xmath301 and @xmath302 are indecomposable whereas @xmath303 and @xmath304 are not .
indeed , @xmath305 and @xmath306 .
[ prop : special - inv - factoring ] each nonzero special invariant is represented uniquely as a product of indecomposable special invariants .
there is a simple algorithm for finding such a factorization ; see section [ sec : special - proofs ] .
indecomposable special invariants will later play the role of ( a subset of ) generators for the cluster structure in @xmath111 .
we call two special invariants _ compatible _ if their product is a single web invariant . this terminology will prove consistent with the notion of compatibility of cluster variables ; cf .
conjecture [ conj : cluster - compatibility ] .
the reader is welcome to check ( this may take some time ) that the four special invariants in figure [ fig : webs61 ] are pairwise compatible .
[ prop : coeff - special ] let @xmath112 be a signature of type @xmath23 , with @xmath253 , @xmath307 , and @xmath308 $ ]
. then there are @xmath104 special invariants compatible with all special invariants .
these @xmath104 invariants are the ( indecomposable ) nonzero invariants of the form @xmath309 .
moreover the product of any of them and any web invariant is a web invariant .
anticipating their future role , we call these special invariants _ coefficient invariants_. we denote the @xmath310-element set of these invariants by @xmath311
. see figure [ fig : webs64 ] .
[ rem : xoxox ] for @xmath312 $ ] , the special invariant @xmath313 factors : @xmath314 . as a result , there are @xmath315 coefficient invariants in this case , namely @xmath316 . in order to write exchange relations for the cluster algebra @xmath111 , we will need certain _
@xmath2-term skein relations _ for special invariants .
[ prop:3-term - skein - special ] let @xmath265 ( respectively @xmath269 ) be distinct boundary vertices , listed in clockwise order .
then @xmath317 it is important to note that identities
hold irrespective of the choice of the signature @xmath112 .
for some of those choices , these identities are illustrated in figure [ fig : webs56 ] .
[ rem : distill ] in each of identities
, some special invariants might be equal to @xmath318 , while others might factor further .
after everything is expressed in terms of indecomposable special invariants , we either get a tautological formula @xmath319 , or else a genuine relation .
we call the latter relation the _ distilled form _ of the original one .
while the distilled relation is always a skein relation of some sort , it does not have to be an instance of ; an example is given in .
[ example : distill ] let @xmath320 $ ] as before .
one instance of equation is @xmath321 .
this does not yield a nontrivial identity : after substituting @xmath322 , @xmath306 , @xmath323 , and @xmath324 , everything cancels out .
a more interesting example is @xmath325 ( cf . ) .
substituting the factorizations @xmath324 and @xmath326 and simplifying , we get @xmath327
in this section , we construct a family of distinguished seeds in ( the quotient field of ) the ring @xmath328 .
these seeds will be used in section [ sec : main - theorem ] to define a cluster algebra structure in @xmath111 .
[ def : z(t ) ] let @xmath22 be a triangulation of our @xmath17-gon by its diagonals .
let @xmath329 to be the collection of special invariants built as follows : * for each diagonal or side @xmath330 in @xmath22 , include @xmath263 and @xmath297 ; * for each triangle @xmath331 in @xmath22 ( here @xmath265 are ordered clockwise ) , include @xmath266 . the _ extended cluster _ @xmath332 consists of all indecomposable special invariants which appear in factorizations of nonzero elements of @xmath329 into indecomposables .
the _ cluster _ @xmath333 consists of all non - coefficient invariants in @xmath332 .
[ th : z(t ) ] for any triangulation @xmath22 as above , the following holds : * the extended cluster @xmath332 contains the entire set @xmath311 of coefficient invariants ; * @xmath332 consists of @xmath334 invariants ; * all special invariants in @xmath332 are pairwise compatible .
[ example : k(t ) ] let @xmath335 $ ]
. then @xmath336 ( see figure [ fig : webs64 ] ) . for the triangulation
@xmath22 shown in figure [ fig : webs60 ] on the left , we have @xmath337 now , @xmath338 and @xmath339 , while several other elements of @xmath329 factor : @xmath340 we conclude that @xmath341 where @xmath342 figure [ fig : webs66 ] shows the webs corresponding to the elements of @xmath343 .
surprisingly , different triangulations may define the same cluster .
[ pr : x(t)=x(t ) ] let @xmath22 and @xmath51 be triangulations which contain the diagonal and coincide outside the quadrilateral @xmath344 .
if the colors of the vertices @xmath345 alternate , then @xmath346 .
[ example : x(t)=x(t ) ] the two triangulations in figure [ fig : webs60 ] yield the same cluster .
( apply proposition [ pr : x(t)=x(t ) ] with @xmath347 . ) to check this directly , refer to example [ example : k(t ) ] , write @xmath348 and use the factorizations @xmath349 , @xmath350 , and @xmath326 . to complete our construction , we need to describe the quivers @xmath351 which , together with the extended clusters @xmath332 , will form the seeds defining a cluster algebra structure in the ring @xmath352 . to put it differently
, we need to write the corresponding exchange relations .
these are obtained from the @xmath2-term skein relations using a couple of additional observations .
let @xmath22 be a triangulation as in definition [ def : z(t ) ] .
if @xmath331 and @xmath353 are triangles in @xmath22 , then by construction the first factor on the left - hand side of each relation is an element of @xmath329 . in order for such a relation ( or rather its distilled form , see remark [ rem : distill ] ) to be a good candidate for an exchange relation out of the extended cluster @xmath332 , we would like the terms on the right - hand side to factor into the elements of @xmath332 . for
, this is true by design ; for , the terms @xmath354 and @xmath267 will only have the requisite property in some special cases .
we say that a side @xmath330 of a triangle @xmath331 is _ exposed _ if it lies on the boundary of the @xmath17-gon ; in other words , @xmath330 is exposed if @xmath355 .
[ pr : thin - triangles ] suppose a triangle @xmath331 has an exposed side .
then @xmath356 ( which element of this set @xmath267 is equal to depends on the signature ) .
[ pr : thin - quads ] let @xmath357 be distinct vertices of the @xmath17-gon .
@xmath358 [ def : exch - rel - t ] continuing in the setting of definition [ def : z(t ) ] , write the following identities : * for each triangle @xmath331 of the triangulation @xmath22 , write formula ; * for each diagonal @xmath359 in @xmath22 separating triangles @xmath331 and @xmath353 : * * write formula ; * * if one of the sides of @xmath331 is exposed , write ; * * if two sides of @xmath331 are exposed , write the appropriate instance of or , if applicable .
it is easy to check , with the help of proposition [ pr : thin - triangles ] , that each of the two monomials on the right - hand side of each of resulting relations will either vanish ( in which case the relation is of no use to us ) or else factor into nontrivial indecomposables . in the latter case , we distill the relation ( see remark [ rem : distill ] ) to obtain a @xmath2-term relation in indecomposable special invariants all of which , with the exception on the second factor on the left , will belong to @xmath332 . to obtain the final list of exchange relations , we should also inspect all instances where we can apply proposition [ pr : x(t)=x(t ) ] to get another triangulation @xmath51 with the same cluster @xmath360 , then check whether applying the above recipe to @xmath51 yields any additional @xmath2-term relations .
a few simple observations help reduce the amount of work required to write the @xmath2-term relations of definition [ def : exch - rel - t ] for a particular choice of @xmath22 and @xmath112 .
for example , if a triangle @xmath331 has an exposed side , then one of the terms on the right - hand side of vanishes , by virtue of proposition [ prop : special=0 ] .
thus , we only need to write relations for the triangles whose all sides are diagonals .
[ prop : special - seeds - exchanges ] for any triangulation @xmath22 of the @xmath17-gon , the procedure described in definition [ def : exch - rel - t ] yields as many relations as there are elements in the cluster @xmath343 .
more precisely , it yields one relation of the form @xmath361 for each @xmath362 ; here @xmath363 are monomials in the elements of @xmath332 .
[ def : q(t ) ] we are now ready to define the _ quiver @xmath351 associated with a triangulation @xmath22_. this quiver is built so that the exchange relations match the @xmath2-term relations obtained via the process described in definition [ def : exch - rel - t ] .
this requirement defines @xmath351 up to simultaneous reversal of direction of all edges incident to any subset of connected components of the mutable part of the quiver .
typically , there will be a single connected component , so that @xmath351 is defined up to a global change of orientation .
it is not hard to see that the choice of a particular incarnation of @xmath351 is inconsequential , as the resulting cluster structure will not depend on this choice . to simplify presentation
, we henceforth use the notation @xmath351 without mentioning its ambiguity .
we continue with the triangulation @xmath22 discussed in example [ example : k(t ) ] ; cf . figure [ fig : webs60 ] on the left .
the only relevant instance of formula is @xmath364 after factoring @xmath365 into indecomposables and substituting @xmath339 ( so as to use consistent notation for the coefficient invariants ) , we obtain @xmath366 among 10 instances of , the nontrivial identities are ( in distilled form ) : @xmath367 j_2 ^ 5 j_7 ^ 4 & = j_{257 } j_4 ^ 5 + j_3 ^ 2 j_7 ^ 5 \qquad \text{(for $ p=2,\ q=3,\ r=5,\ s=7$),}\\[.1 in ] \label{eq : j52*j73 } j_5 ^ 2 j_7 ^ 3 & = j_8 ^ 2 j_1 ^ 7 j_5 ^ 3 + j_{257 } \qquad \text{(for $ p=5,\ q=7,\ r=2,\ s=3$),}\\[.1 in ] \label{eq : j75*j26 } j_7 ^ 5 j_2 ^ 6 & = j_2 ^ 5 j_7 ^ 6 + j_{257}\quad \qquad \text{(for $ p=7,\ q=2,\ r=5,\ s=6$).}\end{aligned}\ ] ] for each of and , there are 7 instances to check .
equation produces one nontrivial relation : @xmath368 equation yields three relations but
all of them replicate those obtained earlier : * for @xmath369 , we get ; * for @xmath370 , we get ; * for @xmath371 , we get .
the rule supplies two equations one of which results in a new nontrivial identity : @xmath372 and there are no instances of . we have obtained 7 exchange relations , one for each element of @xmath343 ( cf . ) except for @xmath302 . to get the missing relation
, we need to replace @xmath22 by the triangulation @xmath51 shown in figure [ fig : webs60 ] on the right .
example [ example : x(t)=x(t ) ] ) .
the only new relation is obtained by applying with @xmath373 .
as explained in example [ example : distill ] , this results in the equation : @xmath374 the corresponding quiver @xmath351 has 16 vertices : 8 frozen , labeled by the coefficient invariants ( see figure [ fig : webs64 ] ) , and 8 mutable , labeled by the elements of the cluster @xmath343 ( see ) . one by one , we draw the edges incident to each mutable vertex using the corresponding exchange relation , so as to match the formula .
this occasionally requires swapping the two terms on the right - hand side . the resulting quiver @xmath351 is shown in figure [ fig : webs65 ] .
it has cluster type @xmath375 .
a general recipe for building the quiver @xmath351 associated with an arbitrary triangulation @xmath22 of the polygon @xmath376 is described in section [ sec : building - a - quiver ] .
additional examples of seeds @xmath377 associated with particular triangulations @xmath22 and signatures @xmath112 can be found in sections [ sec : t - fan ] and [ sec : examples - of - seeds ] .
after all the preparation of the preceding sections , the statement of our main theorem comes as no surprise .
[ th : main ] let @xmath112 be a non - alternating cyclic signature of type @xmath23 , and let @xmath376 be a convex @xmath17-gon with vertices colored according to @xmath112 .
let @xmath22 be a triangulation of @xmath376 by its diagonals .
then the extended cluster @xmath332 and the quiver @xmath351 constructed as described in definitions [ def : z(t ) ] and [ def : q(t ) ] form a seed in the quotient field of the ring of invariants @xmath111 whose associated cluster algebra is @xmath111 : @xmath378 this cluster structure on @xmath111 does not depend on the choice of triangulation @xmath22 .
any element of @xmath111 , and in particular any web invariant , can be expressed as a laurent polynomial in the elements of any extended cluster @xmath332 . for the record , we state below some basic properties of the cluster algebra @xmath111 that are immediate from its construction ( as justified by theorem [ th : main ] ) .
the cluster algebra @xmath111 has the following properties : 1 .
the set of coefficient variables and cluster variables includes all indecomposable special invariants , and in particular all weyl generators . 2 .
the coefficient variables are the coefficient invariants ( cf . proposition [ prop : coeff - special ] ) .
each extended cluster consists of @xmath334 elements ; 4 .
the rank ( the cardinality of each cluster ) is @xmath379 , except for the cases @xmath312 $ ] and @xmath380 $ ] when it is equal to 1 ( cf .
remark [ rem : xoxox ] ) .
let @xmath381 , @xmath382 , @xmath383 $ ] . in this case
, the cluster algebra @xmath111 has type @xmath179 .
it has @xmath243 coefficient variables , namely : @xmath384 ( some of these have alternative presentations , for example @xmath385 . )
the @xmath243 cluster variables in @xmath111 are : @xmath386 they form @xmath243 clusters : @xmath387 the exchange relations are : @xmath388 notice that the coefficient variable @xmath389 does not appear in these relations .
accordingly , the frozen vertex corresponding to @xmath389 is isolated in every quiver of @xmath111 .
one example of such a quiver is shown in figure [ fig : xxxoo ] .
( 80,22)(0,-2 ) ( 0,20)(0,0)@xmath390 ( 20,0)(0,0)@xmath391 ( 20,20)(0,0)@xmath392 ( 40,0)(0,0)@xmath393 ( 40,20)(0,0)@xmath394 ( 60,20)(0,0)@xmath395 ( 80,20)(0,0)@xmath389 ( 16,20)(-1,0)12 ( 56,20)(-1,0)12 ( 24,20)(1,0)12 ( 20,4)(0,1)12 ( 40,16)(0,-1)12 while the cluster structure on @xmath113 defined in theorem [ th : main ] is independent of the choice of triangulation , it does critically depend on the signature @xmath112 .
in fact , fixing the parameters @xmath15 and @xmath14 does not determine the cluster type of @xmath111 , nor whether @xmath111 is of finite or infinite type .
see figure [ fig : boundary - signatures-5678 ] .
lll [ cols="<,<",options="header " , ] for each signature @xmath112 , there are choices of a triangulation @xmath22 for which the construction of the quiver @xmath351 simplifies considerably .
one such choice is presented in section [ sec : t - fan ] , providing a totally explicit rule for determining the ( extended ) cluster type of @xmath111 for any signature @xmath112 .
we next discuss the important special case @xmath396 ( ring of invariants of @xmath14 vectors in ) , or equivalently the case of monochromatic signature @xmath397\ ] ] of type @xmath398 . as mentioned in section [ sec : cluster - algebras ] , cluster structures in the rings @xmath97 , and indeed in homogeneous coordinate rings of arbitrary grassmannians @xmath399 , were described by j. scott @xcite and extensively studied thereafter ( cf . , e.g. , @xcite ) .
[ th : we = scott ] the cluster algebra structure in the ring of invariants @xmath97 ( or equivalently in the homogeneous coordinate ring of the grassmannian @xmath400 ) described in theorem [ th : main ] coincides with the one given by j. scott @xcite . in the grassmannian case @xmath396 , our construction of seeds @xmath377 simplifies considerably , see figure [ fig : webs47 ] . this construction generalizes the one given by scott @xcite for a particular kind of triangulation @xmath22 .
[ example : g3 - 678 ] for @xmath401 , the cluster algebras @xmath402 $ ] are of finite type , so we can list all their generators explicitly .
unsurprisingly , we end up reproducing scott s original description of these generators @xcite in the language of webs .
in fact , scott gave a geometric interpretation of these generators ( for @xmath401 ) which can be seen to match our construction .
the cluster types of the rings @xmath403 $ ] , @xmath401 , are well known to be as follows : @xmath179 for @xmath404 , @xmath405 for @xmath406 , @xmath407 for @xmath408 , and @xmath375 for @xmath409 ( see , e.g. , @xcite ; cf .
figure [ fig : boundary - signatures-5678 ] ) .
the set of generators of each cluster algebra @xmath97 includes the @xmath410 plcker invariants @xmath411 given by tripod webs . among them
there are @xmath14 coefficient variables of the form @xmath412 . for @xmath413
, there are also non - plcker cluster variables .
the simplest among them are the `` hexapod '' invariants given by webs of the kind shown in figure [ fig : webs50 ] on the left .
there are @xmath414 of these hexapods in @xmath97 . for @xmath406 and @xmath408 ,
the tripods and hexapods comprise the entire list of generators . for @xmath409
, there are 24 additional cluster variables , namely 8 ( resp . , 16 ) web invariants of the kind shown in the second ( resp . , third )
picture in figure [ fig : webs50 ] .
cluster algebra structures in rings of invariants @xmath111 behave nicely under two types of embeddings of such rings into one another . to state these results , we will need the following natural definition .
let @xmath152 be a cluster algebra .
take a seed @xmath140 of @xmath152 .
freeze some subset of cluster variables in @xmath141 ( thus , some subset of mutable vertices in @xmath131 ) .
if the resulting quiver has vertices that are not connected by an edge to a mutable vertex , remove some of them ( and remove the corresponding elements from @xmath141 ) to get a new seed @xmath149 .
the cluster algebra @xmath415 obtained in this way is called a _
cluster subalgabra _ of @xmath152 .
[ th : drop - vertex ] let @xmath112 and @xmath416 be two non - alternating signatures such that @xmath416 is obtained from @xmath112 by removing a single symbol .
then the image of @xmath417 under the obvious tautological embedding @xmath418 is a cluster subalgebra of @xmath111 .
theorem [ th : drop - vertex ] seems to be new even in the case of grassmannians ( cf . theorem [ th : we = scott ] ) .
for example , it implies that any web invariant of the form shown in figure [ fig : webs50 ] is a cluster variable in @xmath97 for any @xmath419 , or indeed in any cluster algebra @xmath13 with @xmath413 , respectively @xmath419 .
recall that @xmath1 is a three - dimensional vector space endowed with a volume form .
cross product _ of two vectors @xmath276 is the covector @xmath277 given by @xmath278 , for @xmath279 .
one similarly defines a cross product of two covectors in @xmath21 , which is a vector .
[ th : fork ] let @xmath112 and @xmath416 be two non - alternating signatures such that @xmath416 is obtained from @xmath112 by replacing two consecutive entries of the same color by a single entry of the opposite color . interpreting this operation algebraically as a cross product , consider the corresponding embedding @xmath418 .
then the image of @xmath417 under this embedding is a cluster subalgebra of @xmath111 .
theorems [ th : drop - vertex ] and [ th : fork ] can be restated as saying that if @xmath416 is obtained from @xmath112 by removing an entry , or by replacing @xmath420 or @xmath421 , or more generally by a sequence of such steps , then @xmath417 can be viewed as a cluster subalgebra of @xmath111 .
in particular , in each of those cases the cluster type of @xmath417 can be obtained by taking an induced subgraph inside some quiver for @xmath422 .
one consequence of this hierarchy is that each cluster algebra @xmath422 can be interpreted as a cluster subalgebra of some grassmann cluster algebra @xmath97 . at the level of tensor diagrams , the embeddings in theorems [
th : drop - vertex][th : fork ] correspond to simple transformations near the boundary of our disk . in the case of theorem
[ th : drop - vertex ] , passing from @xmath416 to @xmath112 corresponds to adding an isolated vertex . for the cross product embedding of theorem [
th : fork ] , we replace one boundary vertex by a pair of vertices of the opposite color , simultaneously adding _ forks _ to all edges of a tensor diagram which use this vertex , see figure [ fig : webs17 ] .
theorems [ th : drop - vertex ] and [ th : fork ] can be used to obtain the following corollary .
[ cor : planar - tree ] if a tensor diagram @xmath180 is a planar tree , then the corresponding web invariant @xmath183 $ ] is a cluster or coefficient variable in @xmath422 .
an observant reader may have noticed that our definition of the cluster structure in the ring @xmath111 does not treat the black and white colors in exactly the same way .
the only place where the symmetry is broken is definition [ def : z(t ) ] : the set @xmath329 used to define the extended cluster @xmath332 includes the invariants @xmath266but not @xmath267 .
[ th : swap - colors ] interchanging the two colors in the definition of the extended clusters @xmath332 does not change the resulting cluster structure in the ring @xmath111 . besides swapping the colors
, one can also consider the `` mirror image '' of our main construction .
we expect this version to yield the same result .
reversal of direction ( more specifically , replacing the clockwise direction in the definition of special invariants by the counterclockwise one ) does not affect the cluster structure in @xmath111 .
we do not know of a natural construction of a cluster algebra structure in the ring @xmath111 for the _ alternating _ signature @xmath423 $ ] ; this case seems to be genuinely exceptional .
if @xmath112 is alternating , then the structure one gets by applying the approach of sections [ sec : special - invariants][sec : special - seeds ] is one of a `` generalized cluster algebra '' in which some of the exchange relations have more than two terms on the right - hand side .
in this section , we discuss conjectural links between * the cluster structure in a ring of invariants @xmath111 ( cf . section [ sec : main - theorem ] ) and * kuperberg s basis of this ring formed by the web invariants ( cf . section [ sec : webs ] ) .
the conjectures in this section are intentionally formulated in general terms , so as to suggest their possible extensions to other contexts , including other classical rings of invariants and cluster algebras associated with marked surfaces .
we call a web invariant @xmath132 _ indecomposable _ if it can not be expressed as a product of two web invariants .
( this is likely equivalent to irreducibility of @xmath132 . ) for example , any coefficient variable is an indecomposable web invariant .
[ conj : cluster - vars - are - in - web - basis ] all cluster variables are indecomposable web invariants .
two cluster variables are called _ compatible _ if they belong to the same cluster .
coefficient variables are compatible with each other , and with all cluster variables .
[ conj : cluster - compatibility ] two cluster ( or coefficient ) variables are compatible if and only if their product is a web invariant .
note that by proposition [ prop : coeff - special ] , the product of a coefficient variable and any web invariant ( in particular , by conjecture [ conj : cluster - vars - are - in - web - basis ] , any cluster variable ) is a web invariant
. the simplest illustration of conjecture [ conj : cluster - compatibility ] can be found in figure [ fig : tripod - and - a - stick ] : the invariants @xmath394 and @xmath389 shown there are compatible with each other , and their product is a web invariant shown in figure [ fig : edge - labelings ] . to see more interesting examples , take any two ( compatible ) webs in figure [ fig : webs66 ] and verify that their superposition can be converted into a single web by iterated skein transformations .
we conjecture that much more is true .
define a _ cluster monomial _
@xcite ) as a monomial in the elements of any given extended cluster . for cluster algebras defined by quivers ,
cluster monomials are known to be linearly independent @xcite .
[ conj : cluster - monomials - are - in - web - basis ] all cluster monomials are web invariants .
while the web in figure [ fig : g38](a ) has no internal cycles , the web in ( b ) has one ; still , this web can be transformed into the non - planar tensor diagram ( c ) which has no such cycles .
( in principle , one can always get rid of the cycles ; the challenge is to do so without getting a linear combination of two or more webs . )
@xmath424 @xmath425 + conjecture [ conj : cluster - monomials - are - in - web - basis ] suggests the following property of the web basis .
[ conj : compat - web - invariants ] given a finite collection of distinct web invariants , if the product of any two of them is a web invariant , then so is the product of all of them . in conjecture
[ conj : compat - web - invariants ] , we can not replace `` the product of all of them '' by `` any monomial in them : ''
it is possible to construct a web invariant ( see section [ sec : imaginary - elts ] ) whose square is not a web invariant .
this invariant is _ not _ a cluster monomial , or else its square would provide a counterexample to conjecture [ conj : cluster - monomials - are - in - web - basis ] .
we note that it is enough to prove conjecture [ conj : compat - web - invariants ] for collections of three invariants .
the general case would then follow by induction .
conjecture [ conj : compat - web - invariants ] is inspired by the `` flag property '' of cluster complexes ( conjectural , but proved in many instances ) , and by the analogous property of the dual canonical basis ( cf .
@xcite ) .
we anticipate that all of the above conjectures generalize broadly .
our most optimistic hopes are expressed in conjecture [ conj : a - and - b ] whose statements are patterned after conjectures [ conj : cluster - compatibility][conj : compat - web - invariants ] and proposition [ prop : coeff - special ] .
[ conj : a - and - b ] in any cluster algebra @xmath152 of geometric type , there exists an additive basis @xmath426 with the following properties : 1 .
[ enum : cluster - mon - in - b ] all cluster monomials in @xmath152 lie in @xmath426 .
two cluster variables are compatible if and only if their product lies in @xmath426 .
the product of a coefficient variable and an element of @xmath426 lies in @xmath426 .
[ enum : clique ] if @xmath235 is a finite subset of @xmath426 such that the product of any two distinct elements of @xmath235 lies in @xmath426 , then the product of all elements of @xmath235 lies in @xmath426 .
5 . among all subsets @xmath427 of indecomposable elements satisfying the condition in
, the ones of maximal cardinality are exactly the extended clusters .
let @xmath132 , @xmath139 , @xmath428 , @xmath429 , be elements of @xmath426 satisfying . if the product of any two among these four , except for @xmath430 , lies in @xmath426 , and @xmath132 and @xmath139 are indecomposable , then @xmath132 and @xmath139 are cluster variables , and is an exchange relation .
suppose that an indecomposable element @xmath431 and a cluster variable @xmath139 satisfy a 3-term relation such that @xmath428 , @xmath429 , @xmath432 , and @xmath433 lie in @xmath426 whereas @xmath434 and @xmath435 do not .
then @xmath132 is not a cluster variable .
it is already nontrivial to show the existence of a basis @xmath426 satisfying a small subset of the conditions in conjecture [ conj : a - and - b ] .
for example , the existence of a basis @xmath426 satisfying is equivalent to linear independence of cluster monomials .
perhaps even more important is the problem of determining a cluster structure in a given ring which is in some sense `` compatible '' with a particular distinguished basis @xmath426 . in the case of the web basis
, we used the properties in conjecture [ conj : a - and - b ] as the guiding principles in designing the cluster structure in the rings of invariants @xmath111 .
one can not help but wonder whether this kind of `` reverse engineering '' process can be made algorithmic , or at least axiomatic .
a couple of approaches to this problem are outlined below . at the heart of the matter
lies this question : what distinguishes cluster variables among other indecomposable elements of the basis ? .
] one conjectural answer involves an extension of the notion of compatibility of cluster variables .
let us call a set of distinct indecomposable web invariants a _ clique _ if their product ( variant : the product of any two of them , cf .
conjecture [ conj : compat - web - invariants ] ) is again a web invariant . by conjecture [ conj :
cluster - monomials - are - in - web - basis ] , any extended cluster is a clique . [
conj : largest - clique ] among all cliques , extended clusters are precisely the ones of the largest cardinality .
thus , an indecomposable web invariant @xmath132 is a cluster variable if and only if the largest clique containing @xmath132 has the cardinality of an extended cluster .
while conjecture [ conj : largest - clique ] may be aesthetically pleasing , it is rather impractical as a tool for determining whether a particular basis element is a cluster variable .
it also is of little help in finding the exchange relations , or equivalently the quivers accompanying extended clusters .
for that , we need to move beyond purely multiplicative properties of web invariants ( to put it differently , beyond relations they satisfy ) into the realm of _
3-term relations_. [ conj:3-term - cluster ] suppose that indecomposable web invariants @xmath132 and @xmath139 satisfy the 3-term relation @xmath436 such that @xmath428 , @xmath429 , @xmath434 , @xmath435 , @xmath432 , @xmath433 , and @xmath437 are web invariants .
then @xmath132 and @xmath139 are cluster variables , and is an exchange relation . note that by conjectures [ conj : cluster - vars - are - in - web - basis][conj : cluster - monomials - are - in - web - basis ] , we expect each exchange relation in our cluster algebra to satisfy the conditions in conjecture [ conj:3-term - cluster ] .
there is also a test that can ( conjecturally ) disqualify a web invariant from being a cluster variable .
[ conj:3-term - non - cluster ] suppose that an indecomposable web invariant @xmath132 and a cluster variable @xmath139 satisfy a 3-term relation such that @xmath428 , @xmath429 , @xmath432 , and @xmath433 are web invariants whereas @xmath434 and @xmath435 are not .
then @xmath132 is not a cluster variable . in conjecture [ conj:3-term - non - cluster
] , the requirement that @xmath432 , and @xmath433 be web invariants can not be dropped see section [ sec : fake - exchange ] for a relevant counterexample .
we conjecture that _ any _ cluster structure can be uniquely recovered from a suitably chosen additive basis @xmath426 using appropriate versions of the criteria in conjectures [ conj : largest - clique][conj:3-term - non - cluster ] . from its very inception ,
cluster theory was motivated by the desire to better understand the ( dual ) `` canonical '' bases in the corresponding rings .
recall that for most of the rings @xmath111 the dual canonical basis is _ different _ from the web basis
although the two bases share many important features , see remark [ rem : khovanov - kuperberg ] .
the canonical basis is expected to have strong positivity properties , such as those spelled out in the conjecture below , which has long been part of the cluster algebras folklore .
[ conj : strong - positivity ] in any cluster algebra @xmath152 of geometric type , there is an additive basis @xmath438 which includes the cluster monomials and has nonnegative structure constants .
that is , any product of elements of the basis @xmath438 should have nonnegative coefficients when expanded in the same basis .
this condition suggests the existence of a monoidal categorification , wherein structure constants become tensor product multiplicities .
see @xcite .
the web basis does not always satisfy the conditions of conjecture [ conj : strong - positivity ] : [ prop : negative - structure - const ] for some choices of signature , some structure constants of the web basis are negative .
an example justifying proposition [ prop : negative - structure - const ] is given in section [ sec : negative - structure - const ] .
the _ laurent positivity conjecture _
@xcite predicts that the laurent polynomial expressing any cluster variable in terms of any given seed has positive coefficients .
this conjecture has been proved in many special cases , see in particular @xcite .
it is well known , and easy to see , that conjecture [ conj : strong - positivity ] is _ stronger _ than laurent positivity .
to deduce the latter from the former , multiply a cluster variable by the denominator of its laurent expansion .
the result is a linear combination of cluster monomials . by conjecture [ conj : strong - positivity ] , they all belong to the basis @xmath438 ; moreover the coefficients in this linear combination must be positive , and we are done .
one possible approach to conjecture [ conj : strong - positivity ] is based on the idea advocated in @xcite ( cf.also fock - goncharov2 ) that a canonical basis can potentially be defined by looking at extremal rays in a suitable cone of `` positive elements '' inside the cluster algebra .
let us call an element @xmath132 of a cluster algebra @xmath152 _ positive _ if @xmath132 is expressed as a laurent polynomial with nonnegative coefficients in terms of any extended cluster @xmath141 .
the _ positive cone _
@xmath439 formed by such elements is closed under products .
the laurent positivity conjecture asserts that the cluster monomials belong to this cone .
one hopes that in fact , much more is true ( cf .
@xcite ) .
[ con : positive - cone ] choosing a generator for each extremal ray of the positive cone yields a basis satisfying the conditions in conjecture [ conj : strong - positivity ] .
equivalently , generators of @xmath440 form a linear basis of @xmath152 that includes the cluster monomials .
positivity of structure constants is then automatic .
by definition , a web invariant can be represented by a single planar tensor diagram ( a web ) .
we conjecture that the cluster variables in @xmath111 are distinguished from other web invariants by the property of possessing a particular kind of alternative presentation ; informally speaking , they can be defined by tensor diagrams which are _
trees_. let us explain .
the _ unclasping _ of a tensor diagram @xmath180 is the graph obtained from @xmath180 by replacing each boundary vertex @xmath255 , say of degree @xmath4 , by @xmath4 distinct vertices serving as endpoints of the edges formerly incident to @xmath255 . by a harmless abuse of terminology
, we shall call a tensor diagram @xmath180 whose unclasping has no cycles a _ forest diagram _ ; if moreover the unclasping is connected , we call @xmath180 a _ tree diagram_. we emphasize that such a diagram @xmath180 does _ not _ have to be planar .
[ conj : cluster - variables - are - trees ] a web invariant @xmath132 is a cluster monomial if and only if @xmath441 $ ] for some forest diagram @xmath180 . to illustrate , consider the webs for non - plcker cluster variables in @xmath442 shown in figure [ fig : webs50 ] .
the first one is a tree .
the second one has a @xmath315-cycle but it goes through a boundary vertex .
the third web does have an internal cycle but the alternative presentation of this invariant shown in figure [ fig : g38-nonplanar - tree ] unclasps to a ( non - planar ) tree .
( 70,75)(0,0 ) ( 35,5 ) ( 35,5)(0,1)18 ( 0,40 ) ( 0,40)(3,-2)13 ( 10,60 ) ( 10,60)(3,-2)13 ( 10,60)(3,1)13 ( 10,15 ) ( 10,15)(1,3)4.2 ( 25,40 ) ( 25,40)(2,-3)8.7 ( 25,40)(-1,-1)8.5 ( 25,40)(0,1)8 ( 35,55 ) ( 35,55)(1,0)8 ( 35,55)(-2,-1)8.2 ( 35,55)(-1,1)8.4 ( 35,75 ) ( 35,75)(-1,-1)8.4 ( 45,40 ) ( 45,40)(-2,-3)8.7 ( 45,40)(1,-1)8.5 ( 45,40)(0,1)13 ( 60,15 ) ( 60,15)(-1,3)4.2 ( 60,60 ) ( 60,60)(-3,-1)13 ( 70,40 ) ( 70,40)(-3,-2)13 ( 35,25 ) ( 15,30 ) ( 55,30 ) ( 45,55 ) ( 25,50 ) ( 25,65 ) ( 70,75)(0,0 ) ( 35,5 ) ( 35,5)(0,1)18 ( 0,40 ) ( 0,40)(3,-2)13 ( 10,60 ) ( 10,60)(6,-1)32.8 ( 10,60)(4,1)12.7 ( 10,15 ) ( 10,15)(1,3)4.2 ( 25,40 ) ( 25,40)(2,-3)8.7 ( 25,40)(-1,-1)8.5 ( 25,40)(0,1)21.5 ( 35,75 ) ( 35,75)(-5,-6)8.4 ( 45,40 ) ( 45,40)(-2,-3)8.7 ( 45,40)(1,-1)8.5 ( 45,40)(0,1)12 ( 60,15 ) ( 60,15)(-1,3)4.2 ( 60,60 ) ( 60,60)(-5,-2)13 ( 70,40 ) ( 70,40)(-3,-2)13 ( 35,25 ) ( 15,30 ) ( 55,30 ) ( 45,54.2 ) ( 25,63.7 ) @xmath443 @xmath444 @xmath445 + @xmath445 + a much more complicated example is shown in figure [ fig : webs71 ] . we note that corollary [ cor : planar - tree ] establishes a very special case of conjecture [ conj : cluster - variables - are - trees ] , namely the case when the same tensor diagram is both a web and a tree .
we next describe an algorithm that takes a web invariant @xmath132 as input and conjecturally constructs a forest diagram defining @xmath132 , or else concludes that none exists .
this will require some preparation .
let @xmath180 be a tensor diagram , @xmath446 and @xmath447 its two internal vertices , and @xmath448 and @xmath449 two edges incident to @xmath446 and @xmath447 , respectively .
we call vertices @xmath446 and @xmath447 _ siblings _ of each other ( more precisely , `` siblings away from @xmath448 and @xmath449 '' ) if the following happens . for @xmath450 ,
let @xmath451 denote the subgraph of @xmath180 whose edge set consists of those edges which can be reached from @xmath452 without going along @xmath453 or connecting through a boundary vertex .
( in particular , the edge @xmath453 is not in @xmath451 . )
we then want @xmath454 and @xmath455 to be isomorphic binary trees having the same multisets of leaves on the boundary of the disk .
thus , @xmath446 and @xmath447 are siblings if they are obtained from the same multiset of boundary vertices by the same sequence of taking pairwise joins .
[ def : arborizing - step ] suppose a tensor diagram @xmath180 contains a fragment which is : * a quadrilateral with one vertex on the boundary , or * a four - edge path whose endpoints are siblings of each other , looking away from the edges of the path .
an _ arborizing step _ is the transformation of such a diagram @xmath180 shown in figure [ fig : webs27 ] .
[ lem : arborizing - step ] an arborizing step does not change the value of the invariant defined by a tensor diagram .
the _ arborization algorithm _ takes a tensor diagram ( not necessarily planar ) as input , and applies arborizing steps until unable to do so .
see figure [ fig : webs26 ] .
[ th : arborization - confluent ] the arborization algorithm is confluent .
that is , its output does not depend on the choice of an arborizing step made at each stage .
we expect the arborization algorithm to cover all cases in which a web invariant can be given by a forest diagram : [ conj : lift1 ] suppose a web invariant @xmath132 can be defined by a forest diagram @xmath180 that can not be arborized further
. then @xmath180 is obtained from the web defining @xmath132 via the arborization algorithm .
conjecture [ conj : lift1 ] suggests the following enhancement of conjecture [ conj : cluster - variables - are - trees ] .
[ conj : lift2 ] a web invariant @xmath132 is a cluster or coefficient variable ( respectively , a cluster monomial ) if and only if the arborization algorithm applied to the web defining @xmath132 outputs a tree diagram ( respectively , a forest diagram ) .
if a web invariant arborizes to a tree diagram , then we expect it to be a cluster variable ( cf .
conjecture [ conj : lift2 ] ) , so its powers which are cluster monomials should be web invariants as well ( cf . conjecture [ conj : cluster - monomials - are - in - web - basis ] ) .
theorem [ th : powers - arborizable ] below confirms this expectation , thereby providing indirect support for the aforementioned conjectures .
[ th : powers - arborizable ] let @xmath132 be a web invariant defined by a web which arborizes to a tree diagram via the arborization algorithm .
then any power of @xmath132 is a web invariant .
we actually do a bit more : under the assumptions of theorem [ th : powers - arborizable ] , we explicitly describe the web that defines the power @xmath456 of an arborizable web invariant @xmath132 .
[ def : thickening ] let @xmath4 be a positive integer , and @xmath271 a web .
the _ @xmath4-thickening _ of @xmath271 is obtained as follows : * replace each internal vertex of @xmath271 by a `` honeycomb '' fragment @xmath457 shown in figure [ fig : webs29 ] ( boundary vertices stay put ) ; * replace each edge of @xmath271 by a @xmath4-tuple of edges connecting the corresponding honeycombs and/or boundary vertices .
an example is shown in figure [ fig : webs30 ] .
[ th : thickening ] let @xmath132 be a web invariant defined by a web @xmath271 that arborizes to a tree diagram via the arborization algorithm . then each power @xmath456 is a web invariant defined by the @xmath4-thickening of @xmath271 .
in this section , we present examples of web invariants possessing various notable properties . by conjecture [ conj :
cluster - vars - are - in - web - basis ] , the set of indecomposable web invariants includes all cluster variables .
which other web invariants does it include ?
conjecture [ conj : lift2 ] suggests an approach to constructing such invariants : if a web does not arborize to a tree diagram , then the corresponding invariant is not a cluster variable .
a couple of examples are shown in figure [ fig : webs34 ] . the converse to theorem [
th : thickening ] ( or theorem [ th : powers - arborizable ] ) is false : there are lots of web invariants which can not be represented by a tree diagram , yet all their powers are web invariants themselves ; moreover they can be obtained by the thickening procedure .
the simplest example is shown in figure [ fig : webs32 ] .
the unclasping of the second web in figure [ fig : webs32 ] yields the minimal counterexample of m. khovanov and g. kuperberg ( * ? ? ?
* theorem 4 ) .
the square of the web invariant shown in figure [ fig : webs33 ] is not a web invariant ; the exact formula is given in figure [ fig : webs70 ] .
this phenomenon parallels the existence of `` imaginary '' elements in dual canonical bases , first discovered by b. leclerc @xcite .
the identity presented in figure [ fig : webs68 ] shows that in conjecture [ conj:3-term - non - cluster ] , one can not drop the requirement for @xmath432 , and @xmath433 to be web invariants .
figure [ fig : webs69 ] presents an instance where a particular structure constant for the web basis is negative .
this example shows that even the product of two _ cluster variables _ may expand in the web basis with coefficients some of which are negative .
this project has intimate connections with incidence geometry .
the equation @xmath183=0 $ ] can be viewed as a condition imposed on a collection @xmath458 of @xmath95 points in @xmath459 .
( in the simplest instance , a tripod encodes collinearity of three points . ) to understand the web basis , we would like to interpret each condition @xmath183=0 $ ] in terms of the moduli space of configurations of points and lines which contain @xmath458 and have a prescribed _ incidence graph _ @xmath460 .
it is easy to see that if @xmath180 is a tree , then @xmath461 works , recovering the interpretations of the cluster variables in @xmath462 ( @xmath463 ) found by j. scott @xcite . for diagrams with a single cycle
, such a description can be obtained from the recent work of j. richter - gebert and p. lebmeir @xcite .
it is routine to examine the cases and check that any tensor diagram representing a special invariant `` planarizes '' ( via repeated application of skein relations ) into a single non - elliptic web .
the key relation used in these verifications is shown in figure [ fig : basic - step ] .
an example is given in figure [ fig : special - planarize ] .
the factorization of a nonzero special invariant into indecomposables can be obtained by repeated application of the following rules ( cf . figure [ fig : caterpillars ] and definition [ def : special - inv ] ) : 1 .
[ enum : fact - step-1 ] if @xmath255 is white and @xmath258 is black , then @xmath464 and @xmath465 ; 2 .
[ enum : fact - step-2 ] if @xmath255 is white , @xmath259 is black , and @xmath466 is white , then @xmath467 ; 3 .
[ enum : fact - step-3 ] if @xmath255 is white and @xmath259 is black , then @xmath468 ; 4 .
[ enum : fact - step-4 ] if @xmath255 is white , @xmath259 is black , and @xmath466 is white , then @xmath469 ; 5 .
[ enum : fact - step-5 ] if the colors are reversed ( i.e. , @xmath255 is black , etc . ) , then switch the colors and swap subscripts with superscripts in rules above .
an example is given in figure [ fig : webs53 ] .
it is tedious but straightforward to check that this process yields a unique factorization , and that no other collections of special invariants multiply to a single special invariant .
first , we need to prove that a nonzero special invariant not on our list is not compatible with some other special invariant .
this is checked on a case by case basis . for an invariant of the form @xmath263
, one finds an incompatible special invariant @xmath470 such that the diagonals @xmath330 and @xmath471 cross each other .
similarly , if our invariant is @xmath266 or @xmath267 , one can find an incompatible special invariant @xmath472 such that the diagonal @xmath473 crosses the triangle @xmath331 .
( these are special cases of failure of _ weak separation _ , cf .
special invariants @xmath270 are handled analogously .
another claim to check is that a nonzero invariant of the form @xmath309 is compatible with any web invariant .
coefficient invariants come in two flavors : short and long , see figure [ fig : webs54 ] .
the short ones are obviously compatible with any web invariant .
the relevant calculation for a long invariant is shown in figure [ fig : webs55 ] .
applying it around every vertex establishes the claim .
each identity can be obtained by repeated application of skein relations .
a key role is played by the relation shown in figure [ fig : basic - step ] . without loss of generality
, we can assume that side @xmath474 is exposed , so that @xmath475 .
one then checks that @xmath476 if @xmath238 is white , and @xmath477 if @xmath238 is black .
if @xmath255 is white and @xmath259 is black , then @xmath478 .
relation then gives @xmath479 if @xmath255 is black and @xmath259 is white , then @xmath480 , and gives @xmath481
denote @xmath482 .
since the theorem can be verified by direct calculation for @xmath483 , we assume that @xmath484 from now on . by proposition [ prop : coeff - special ] , each side of the @xmath5-gon @xmath376 produces one nonzero special invariant which by construction belongs to @xmath332 ; those are exactly the coefficient invariants .
define the _ length _ of a diagonal @xmath330 in @xmath376 as the number @xmath485 ; this is the length of the shortest path from @xmath255 to @xmath238 along the perimeter of @xmath376 .
[ lem : decomp ] assume that @xmath484 .
let @xmath486 denote the interior of @xmath376 with all diagonals of length @xmath173 in @xmath22 removed .
the set @xmath486 uniquely decomposes into a disjoint union of `` fundamental regions '' of types @xmath487 shown in figure [ fig : webs73 ] .
an example is shown in figure [ fig : webs74 ] .
we note that the lemma fails for @xmath488 if @xmath22 contains a diagonal of length @xmath2 .
( the two regions of type @xmath489 are not disjoint as they share this diagonal . )
special invariants associated to each fundamental region contribute the following number of non - coefficient indecomposable factors to the cluster @xmath343 : @xmath490 the contributions of different fundamental regions are disjoint . to clarify , if a region @xmath16 does not include a ( dashed ) diagonal @xmath330 lying on its boundary , then the contributions of @xmath16 exclude all indecomposable factors appearing in @xmath263 and @xmath297 . case by case consideration depending on the signature . one typical example involving a region of type @xmath489 ( with two possible colorings )
is shown in figure [ fig : webs75 ] .
we continue with the proof of theorem [ th : z(t ) ] .
it is a simple counting exercise to check that the total contribution from all fundamental regions is exactly @xmath491 non - coefficient variables , as claimed . for example , in figure [ fig : webs74 ] we get @xmath492 it remains to check that the special invariants in @xmath332 are pairwise compatible . for invariants coming the same fundamental region ,
this is done by direct case by case inspection . for invariants from different regions ,
the argument goes as follows .
the `` tails '' of the special invariants ( cf .
figure [ fig : caterpillars ] ) do not create any obstructions to compatibility : just iterate the basic planarizing steps of figure [ fig : basic - step ] .
the remaining pieces lie in different regions and thus do not intersect .
the proof reduces to a direct verification comparing triangulations @xmath22 and @xmath51 which differ by a flip that switches the diagonals @xmath493 and @xmath494 .
a typical example is shown in figure [ fig : webs15 ] .
the proof is a case by case verification , organized as follows . for each region @xmath16
, the variables ( i.e. , indecomposable factors ) contributed by @xmath16 are exchanged using only the variables defined within @xmath16 and/or several adjacent regions . specifically : * regions of type @xmath495 do not contribute any variables , so there is nothing to check ; * for regions of type @xmath235 ( resp . ,
@xmath496 , @xmath227 ) , also consider two ( resp .
, three , one ) adjacent region(s ) of types @xmath180 or @xmath489 ; * for regions of type @xmath180 ( resp . , @xmath489 ) , also consider two ( resp . , one ) adjacent region(s ) of types @xmath495 , @xmath235 , @xmath496 or @xmath227 , as well as the regions of types @xmath180 or @xmath489 which are adjacent to those .
in each of these cases , the proof consists of a local verification ( sometimes tedious but always straightforward ) ranging over a finite list of possible patterns .
in this section , we provide a blueprint for building the quiver @xmath351 associated with an arbitrary triangulation @xmath22 of the polygon @xmath376 .
as mentioned earlier , a detailed description for a particular choice of @xmath22 can be found in section [ sec : t - fan ] . the simplest part of the recipe concerns the portions of @xmath351 coming from triangles @xmath331 in @xmath22 which have no exposed sides .
we draw the vertex of @xmath351 representing the special invariant @xmath266 inside the triangle @xmath331 .
the vertices representing cluster variables coming from @xmath263 and @xmath297 are placed on the diagonal @xmath330 ( and similarly for @xmath474 and @xmath359 ) , with the former closer to @xmath238 , and the latter closer to @xmath255 .
see figure [ fig : webs43 ] . the above recipe may require adjustments if some of the sides of the triangle @xmath331 are too short , so that the corresponding special invariants factor ; cf . , e.g. , figure [ fig : webs15 ] . for triangles with one or two sides on the boundary of @xmath376
, the basic principles remain the same , but the recipe changes somewhat .
figures [ fig : webs11 ] and [ fig : webs12 ] treat `` generic '' cases of one and two exposed sides , respectively . in a few exceptional cases ,
additional adjustments have to be made to the construction of the quiver @xmath351 and/or to the rules of assigning special invariants to its vertices .
these adjustments are caused by nontrivial factoring of the special invariants involved .
one such case is illustrated in figure [ fig : webs14 ] , which is in turn a special case of the pattern shown in figure [ fig : webs12 ] on the upper right .
another example is shown in figure [ fig : webs15 ] .
there are only a finite number of such exceptional situations , all of them arising when some sides of the relevant triangle(s ) are short enough to make certain special invariants factor nontrivially .
we refrain from exhaustively describing all these exceptional cases and the corresponding quiver - building instructions .
theorem [ th : main ] will follow once we establish claims below : 1 . [
item : alg - indep ] the elements of @xmath332 are algebraically independent ; 2 . [
item : mut - flips ] the seeds @xmath377 are mutation - equivalent ; 3 . [
item : irreducible ] each indecomposable special invariant is an irreducible element of @xmath111 .
here is why .
claim means that @xmath377 is a seed .
claim means that the cluster algebra @xmath497 does not depend on @xmath22 .
claim , in combination with lemma [ lem : properties - of - rabv ] , corollary [ cor : cluster - criterion ] , claim , the first fundamental theorem of invariant theory , and the fact that all weyl generators show up in extended clusters @xmath332 , implies that @xmath498 .
since all exchange relations from the special seeds @xmath377 involve exclusively special invariants , claim implies that the cluster variables appearing in seeds adjacent to the special ones are irreducible . the bulk of the work ( most of it rather tedious ) goes into establishing claim . since any two triangulations of the polygon @xmath376 can be connected by a sequence of _ flips _
( each flip replacing a single diagonal by another one ) , it suffices to show that for two triangulations that differ by a flip , the corresponding seeds are related by mutations . typically , the required number of mutations is four , although in some cases it can be three , two , one , and even zero , see proposition [ pr : x(t)=x(t ) ] .
the verification of the claim is done on a case by case basis , depending on the color pattern of the vertices involved ; as before , there are several exceptional cases where closeness of vertices results in nontrivial factorization of the corresponding special invariant . once again , we do not list all cases exhaustively , presenting instead a couple of `` generic '' cases which illustrate the checks one needs to perform . the general rule ( barring aforementioned exceptions )
is as follows .
suppose that triangulations @xmath22 and @xmath51 are related by a flip that replaces diagonal @xmath359 by diagonal @xmath499 ; thus @xmath22 has triangles @xmath331 and @xmath500 while @xmath51 has triangles @xmath501 and @xmath502 . then the seed @xmath503 is obtained from @xmath377 by a sequence of four mutations : * replace @xmath504 by @xmath505 , and replace @xmath506 by @xmath507 ( these two mutations commute ) ; * replace @xmath508 by @xmath509 , and replace @xmath266 by @xmath510 ( these two mutations commute ) .
the corresponding exchange relations are of the form . figures [ fig : webs10 ] and [ fig : webs8 ] illustrate these sequences of mutations for two different color patterns . in section [ sec : other - proofs ]
( see the proof of theorem [ th : we = scott ] ) , we examine exceptional instances of mutation sequences associated with diagonal flips in the special case of a monochromatic signature ( equivalently , the case of a grassmannian @xmath511 ) .
we begin by noting that @xmath512 has the same dimension as @xmath513 , the affine cone over the grassmannian @xmath514 ; the latter dimension is equal to @xmath515 . to prove the equality of dimensions , assume without loss of generality that @xmath516 , and use the @xmath0-equivariant birational isomorphism that sends an @xmath15-tuple of covectors @xmath517 to the @xmath15-tuple of vectors @xmath518 . since we know from that the @xmath519-tuples @xmath332 ( cf
. theorem [ th : z(t ) ] ) are birationally related to each other , and collectively generate the field of fractions of @xmath352 ( as they contain all weyl generators ) , it follows that each of these tuples is algebraically independent .
here we prove claim , the irreducibility of indecomposable special invariants .
incidentally , ( * ? ? ?
* theorem 3.17 ) asserts that @xmath114 is irreducible ( as an element of @xmath111 ) if and only if @xmath105 is an irreducible @xmath0-invariant polynomial .
that is , if @xmath105 factors nontrivially in the polynomial ring @xmath520 $ ] , then the factors must be @xmath0-invariant . we do not need to rely on this result .
[ lem : antis ] assume that an invariant @xmath521 is antisymmetric and linear with respect to its arguments @xmath522 and @xmath523 which have the same variance ( i.e. , are both contravariant or both covariant )
. then @xmath168 can be expressed as a linear combination of web invariants defined by webs which have a fork between vertices @xmath255 and @xmath259 .
we can write @xmath168 as a linear combination of web invariants @xmath524 where the web representing each @xmath525 has a single edge incident to @xmath255 ( resp . , to @xmath259 ) .
let us attach a crossing to this web at vertices @xmath255 and @xmath259 to get an invariant @xmath526 .
since @xmath527 is obtained from @xmath168 by interchanging @xmath522 and @xmath523 , and since @xmath168 is antisymmetric in these arguments , it follows that @xmath528 . on the other hand ,
applying the skein relation at the crossing gives @xmath529 , where @xmath530 has a fork between @xmath255 and @xmath259 .
taking linear combinations , we get @xmath531 , implying @xmath532 , an expansion of the desired kind .
[ lem : plantreeirr ] a web invariant defined by a planar tree is irreducible .
note that by corollary [ cor : planar - tree ] , such an invariant is a cluster or coefficient variable in @xmath422 . by (
* theorem 1.3 ) , every cluster variable in _ any _ cluster algebra is irreducible . at this point , we can not of course rely on these statements , as we are still in the process of proving that @xmath422 is a cluster algebra .
we proceed by induction on the size of the tree . in the base cases of weyl generators ,
irreducibility is well known ( and easy to prove ) .
suppose a web invariant @xmath168 defined by a planar tree @xmath22 has a nontrivial factorization @xmath533 .
it is easy to see that @xmath22 must have a fork , say between vertices @xmath255 and @xmath238 ( of the same color ) associated with ( co)vectors @xmath522 and @xmath534 , respectively . without loss of generality , we may assume that @xmath535 . as @xmath168 is multi - homogeneous ,
so must be @xmath536 and @xmath537
. moreover @xmath168 is multilinear , implying the dichotomy : either one of the factors @xmath538 depends ( linearly ) on both @xmath522 and @xmath534 while the other depends on neither , or else one of the factors depends on @xmath522 but not @xmath534 while another depends on @xmath534 but not @xmath522 .
the latter option is ruled out by the fact that @xmath168 is antisymmetric in @xmath522 and @xmath534 ( because of the fork ) , so it must vanish if we substitute @xmath539 . on the other hand , neither factor
@xmath538 vanishes under this substitution , and @xmath352 is a domain .
so @xmath522 and @xmath534 appear in the same factor , say @xmath536 .
then @xmath537 does not depend on @xmath522 and @xmath534 .
hence @xmath536 is antisymmetric in @xmath522 and @xmath534 .
applying lemma [ lem : antis ] , we express @xmath536 as a linear combination of web invariants whose webs have a fork between @xmath255 and @xmath355 .
thus , @xmath522 and @xmath534 enter the identity @xmath533 exclusively through their cross product @xmath540 .
this yields a nontrivial factorization of an invariant defined by a smaller planar tree , contradicting the induction assumption .
we are now ready to prove that any indecomposable special invariant @xmath168 is irreducible .
we describe the general idea of the proof , omitting some details .
the proof is by induction on the number of internal vertices in the tree diagram defining @xmath168 .
let @xmath541 denote the set of indices appearing in the original notation for @xmath168 ( cf . definition [ def : special - inv ] ) .
thus the cardinality of @xmath495 is @xmath173 , @xmath2 , or @xmath204 .
if @xmath168 is a planar tree , the statement reduces to lemma [ lem : plantreeirr ] .
otherwise the `` tail '' involved in building the proxy for some of the vertices in @xmath495 ( cf .
figure [ fig : caterpillars ] ) is long enough to cover another such vertex .
it is not hard to see that we can find two consecutive vertices @xmath542 ( going clockwise ) such that the boundary segment between them is covered by only one such tail .
furthermore , for @xmath168 to be indecomposable , this segment has to be sufficiently long ( cf .
figures [ fig : webs61][fig : webs62 ] to see why ) .
assume without loss of generality that vertex @xmath255 is white ; then @xmath259 is black , @xmath466 is white , etc .
we then define invariants @xmath543 and @xmath544 as follows .
( consult figure [ fig : webs72 ] . ) for @xmath543 , attach an outside fork to @xmath259 and identify one of its endpoints with @xmath255 .
( this changes the color of the entry @xmath259 of the signature . ) for @xmath544 , attach an outside fork to @xmath255 and identify one of its endpoints with @xmath259 .
( this changes the color of the @xmath255th entry . )
thus the invariants @xmath545 and @xmath546 are obtained from @xmath547 via the substitutions @xmath548 suppose @xmath168 has a nontrivial factorization @xmath549 ; it then specializes into factorizations @xmath550 and @xmath551 . on the other hand , @xmath543 and @xmath544 factor
as shown in figure [ fig : webs72 ] ( bottom row ) , one of the factors being given by a planar tree ( the cut - off tail ) and another factor being a special invariant of the same kind as @xmath168 .
the former factor is irreducible by lemma [ lem : plantreeirr ] while the latter is irreducible by the induction assumption ( here we actually need some kind of explicit description of which special invariants are indecomposable ) . since our rings of invariants are unique factorization domains , we conclude that factorizations @xmath550 and @xmath551 must coincide with the ones shown in figure [ fig : webs72 ] . since @xmath168 is linear in @xmath523 , either @xmath172 or @xmath552 ( say @xmath172 ) does not depend on @xmath523 . thus @xmath172 is unaffected by the substitution , and @xmath553 does not depend on @xmath554 .
consequently , @xmath172 must be the invariant given by the cut - off tail , shown in red in the lower - left corner of figure [ fig : webs72 ] .
similarly , one of the factors @xmath172 and @xmath552 does not depend on @xmath522 , is unaffected by the substitution , and appears in the factorization of @xmath544 shown in figure [ fig : webs72 ] in the lower - right corner ; this must be the invariant shown in blue ( and it must be @xmath552 ) . in conclusion , the factors @xmath172 and @xmath552 must match the ones shown in the center of the bottom row .
but the product of these two special invariants yields a sum of two terms , one of them being @xmath168 .
the second term does not vanish because the tail is long enough .
this completes the proof of theorem [ th : main ] .
it is straightforward to verify that the construction described in @xcite arises as a special case of our setup for the `` zig - zag '' triangulation that includes all diagonals of the form @xmath555 and @xmath556 . for the sake of completeness
, we include a case - by - case verification that in the case of monochromatic signature , the seeds @xmath557 related by a single flip are indeed obtained from each other by mutations .
we stated this claim in section [ sec : proof - main ] for an arbitrary signature , but did not conduct an exhaustive examination of all possible non - generic cases .
such an examination ( for @xmath396 ) is presented in figures [ fig : webs49b][fig : webs48b ] .
our goal is to identify the cluster structure after dropping a boundary vertex with a part of the original cluster structure .
there are several cases to consider .
figure [ fig : webs20 ] explains how to handle the case where the color of the dropped vertex differs from both of its neighbors . other cases are treated in a similar way .
we follow the pattern of the proof of theorem [ th : drop - vertex ] , examining one characteristic example ( see figure [ fig : webs19 ] ) instead of conducting a formal case by case analysis .
the size of this example is large enough to demonstrate what happens in general .
any planar tree can be grown by repeated application of the operations of adding a fork and dropping a vertex .
make sure to never pass through an alternating signature ; this can indeed be done .
we omit the details .
the construction of an extended cluster @xmath332 can be carried out verbatim with the black and white colors swapped .
let us denote the resulting set by @xmath558 .
note that the definition of a proxy vertex remains the same .
the only difference between the two constructions is that for each triangle @xmath331 in @xmath22 , we use ( the factors of ) the special invariant @xmath266 to build @xmath332 whereas the construction of @xmath558 involves @xmath267 . in each instance where the latter choice yields an indecomposable factor that does not come from any special invariant contributing to @xmath332
, one can use the relation to mutate between the two seeds .
this can be verified on a case by case basis , treating each fundamental region separately .
a skein relation yields the identity shown in figure [ fig : webs27b ] .
it remains to note that the second term vanishes as it is both symmetric and antisymmetric with respect to the ( co)vectors corresponding to the two sibling vertices .
the proof is a straightforward application of the diamond lemma .
the five possible nontrivial diamonds are shown in figure [ fig : webs28 ] .
the idea of the proof is as follows .
assume that @xmath271 arborizes to a tree diagram @xmath180 .
let us reverse the arborization process : starting with @xmath180 , we step by step `` planarize '' it ( cf .
figure [ fig : basic - step ] ) , each time creating a four - edge fragment of one of the two kinds described in definition [ def : arborizing - step ] .
now make a tensor diagram for @xmath456 by bundling together @xmath4 copies of @xmath180 , and apply a planarization procedure following the same steps , each time planarizing all @xmath4 copies . at the end of this process
, we are going to obtain the @xmath4-thickening of @xmath271 , thereby proving the theorem .
we illustrate how this works using the example in figure [ fig : webs31 ] .
the left picture shows an intermediate stage of the planarization process , with @xmath7 .
the six white vertices correspond to two sibling vertices in @xmath180 .
remember that we are transforming our tensor diagram from the inside out , i.e. , moving towards the boundary .
the portions attached underneath those six vertices all look identical : they are patterned after the same subtree of @xmath180 .
we proceed in two stages as shown in the figure . at the first stage ,
we planarize around the two triples of black vertices , using skein relations ( including the square move ) .
each time , all terms but one vanish because they can be seen to be both symmetric and antisymmetric with respect to a pair of arguments . at the second stage , we planarize the intersection of @xmath4-thickenings of two edges of @xmath180 using a similar calculation .
let us also point out why it is crucial for this argument that the diagram @xmath180 is a tree .
we need to start the process of reverse arborization from some internal vertex . in a tree
, one can choose either of the three directions and have all @xmath4 copies of this vertex have the same origin in this direction .
this allows to planarize the part of the @xmath4 copies of the original web around this vertex ( as it was just described ) .
if @xmath180 is not a tree however , copies of the vertices belonging to cycles do not share a common origin , and thus the argument fails .
the cluster structure in the ring of invariants @xmath111 is in principle determined by a single initial seed @xmath377 .
consequently , one may be interested in identifying a particular choice of a triangulation @xmath22 for which the seed @xmath377 has a simpler and more explicit description than the general case as presented in sections [ sec : special - invariants][sec : special - seeds ] .
one such choice is discussed in this section .
denote @xmath559 .
since the signature @xmath112 is non - alternating , we can assume without loss of generality that the vertices @xmath203 and @xmath173 are black .
( if we can only find two adjacent white vertices , then use the same recipe with the colors swapped , cf .
theorem [ th : swap - colors ] . )
the case @xmath396 has been covered in figure [ fig : webs47 ] .
so let @xmath112 be non - monochromatic .
without loss of generality , we can assume that the vertex @xmath5 is white .
let @xmath560 be the triangulation of the @xmath5-gon @xmath376 obtained by drawing all diagonals with an endpoint at the vertex @xmath203 .
the quiver @xmath561 is constructed as follows . 1 .
place two ( mutable ) vertices of @xmath561 on each diagonal of @xmath560 .
place one frozen vertex on each side of @xmath376 .
2 . fill each triangle in @xmath560 with arrows as shown in figure [ fig : webs44 ] .
the choice of a pattern is determined by the color of the right endpoint of the base side .
3 . assemble a quiver from these pieces , removing @xmath173-cycles if necessary .
4 . on the diagonal @xmath562 , freeze the vertex closer to @xmath203 , and identify it with the frozen vertex @xmath563 .
5 . on the diagonal @xmath564 ,
freeze one vertex , namely the one closer to @xmath203 ( resp .
, @xmath565 ) if @xmath565 is white ( resp . ,
identify it with @xmath566 if @xmath567 is white , and with @xmath568 if it is black .
add arrows connecting the mutable vertices on diagonals @xmath562 and
@xmath564 to the frozen vertices @xmath563 , @xmath569 , @xmath566 , and @xmath568 as shown in figure [ fig : webs44 ] on the right . remove 2-cycles if necessary ; this will happen if vertex @xmath2 is black .
[ enum : except - q ] if @xmath567 is black , replace the portions of the resulting quiver inside the pentagon @xmath570 as shown in figure [ fig : webs45 ] .
figure [ fig : webs46 ] shows an example of applying this construction .
when the vertex @xmath567 is white , the above construction of the quiver @xmath561 simplifies substantially as we do not have to make the adjustments described in step .
if one is only interested in the cluster type of @xmath111 , the description simplifies further : let @xmath571 $ ] , @xmath572 , be a signature such that @xmath573 and @xmath574 .
let @xmath560 be the triangulation of @xmath376 by the diagonals incident to the vertex @xmath203 .
then the mutable part of the quiver @xmath561 is obtained by taking the vertices @xmath575 and @xmath576 together with the following edges : * for @xmath577 , draw an edge @xmath578 ; * for @xmath579 , draw an edge @xmath580 ; * for @xmath581 , if @xmath582 , then draw an edge @xmath583 ; * for @xmath581 , if @xmath584 , then draw an edge @xmath585 ; * for @xmath586 , if @xmath582 , then draw an edge @xmath587 ; * for @xmath586 , if @xmath584 , then draw an edge @xmath588 .
99 c. geiss , b. leclerc , and j. schrer , preprojective algebras and cluster algebras , _ in : _ trends in representation theory of algebras and related topics , 253283 , _ ems ser .
congr . rep .
soc . , zrich , 2008 .
b. leclerc and a. zelevinsky , quasicommuting families of quantum plcker coordinates , in : kirillov s seminar on representation theory , in : _ amer . math .
( 2 ) , vol . 181 _ , amer .
soc . , providence , ri , 1998 , 85108 . v. l. popov and e. b. vinberg , _ invariant theory _ , in : _ algebraic geometry .
iv , encyclopaedia of mathematical sciences , vol . 55 _ , springer - verlag , berlin , 1994 , 123284 .
a. postnikov , d. speyer , and l. williams , matching polytopes , toric geometry , and the totally non - negative grassmannian , _ j. algebraic combin .
_ * 30 * ( 2009 ) , 173191 . | the rings of @xmath0 invariants of configurations of vectors and linear forms in a finite - dimensional complex vector space @xmath1 were explicitly described by hermann weyl in the 1930s .
we show that when @xmath1 is @xmath2-dimensional , each of these rings carries a natural cluster algebra structure ( typically , many of them ) whose cluster variables include weyl s generators . we describe and explore these cluster structures using the combinatorial machinery of tensor diagrams .
a key role is played by the web bases introduced by g. kuperberg . |
cartographic maps are an important tool for exploring , analyzing and communicating data in their geographic context .
effective maps show their main information as prominently as possible .
map elements that are not relevant to the map s purpose are abstracted or fully omitted . in _
schematic maps _ , the abstraction is taken to `` extreme '' levels , representing complex geographic elements with only a few line segments .
in addition to highlighting the main aspects of the map , they are useful to avoid an `` illusion of accuracy '' @xcite which may arise when showing data on a detailed map : the schematic appearance acts as a visual cue of distortion , imprecision or uncertainty .
however , the low complexity must be balanced with recognizability .
correct topological relations and resemblance hence play a key role in schematization .
schematic maps also tend to be stylized by constraining the permitted geometry .
orientations of line segments are often restricted to a small set @xmath4 , so called _
@xmath4-oriented schematization_. the prototypical example is a schematic transit map ( e.g. the london tube map ) , in which all segments are horizontal , vertical or a @xmath5-degree diagonal .
to be schematized ( switzerland ) .
( b ) a @xmath4-oriented ( rectilinear ) graph @xmath2 is placed on @xmath1 .
( c ) a _ simple _ polygon @xmath0 with its boundary constrained to @xmath2 that resembles @xmath1 .
( d ) shape @xmath0 is a @xmath4-oriented schematization of the input . ]
a central problem in schematization is the following : given a simple polygon @xmath1 , compute a simple @xmath4-oriented polygon @xmath0 with low complexity and high resemblance to @xmath1 .
typically , one is constrained to optimize the other .
formalizing `` high resemblance '' requires the use of similarity measures , each having its own benefits and weaknesses @xcite . in this paper
we investigate a discretized approach to schematization . to this end
we overlay a @xmath4-oriented plane graph @xmath2 on @xmath1 , and require the boundary of @xmath0 to coincide with a simple cycle in @xmath2 ( fig .
[ fig : idea ] ) .
though it restricts the solution space , this approach also offers some benefits .
* the graph can easily model a variety of constraints , possibly varying over the map , or mixing with other types of geometry , such as circular arcs .
* discretization promotes the use of collinear edges and provides a uniformity of edge lengths .
this provides a stronger sense of schematization and a more coherent `` look and feel '' when schematizing multiple shapes . *
combined with the simplicity constraint , the graph enforces a minimal width for narrow strips in @xmath1 , avoiding an undesired visual collapse ( see fig . [ fig : exaggeration ] ) .
finally , discretization may be necessary due to the intended use of the schematic shape .
examples include computing tile - based cartograms @xcite and deciding for a grid map @xcite on a connected set of cells that resemble the complete region .
these use cases require us to bound not ( only ) the bends of the result , but the number or total size of the enclosed faces .
this is also relevant in the context of area - preserving schematization @xcite .
exaggerates the narrow strip .
computed using techniques described in @xcite .
( b ) result of algorithm by buchin et al .
@xcite contains a visual collapse . ]
we consider two discretized approaches , as outlined below .
note that we focus on grid graphs ( plane graphs with horizontal and vertical edges only ) without constraining the complexity of the result .
refer to section [ sec : prelims ] for precise definitions used in this paper .
the first approach aims to find a simple cycle in @xmath2 , quantifying resemblance via the frchet distance ( @xmath6 ) .
this leads us to the problem statement below .
[ prob : smm ] let @xmath2 be a partial grid graph , let @xmath1 be a simple polygon and let @xmath7 . decide whether a simple cycle @xmath3 in @xmath2 exists such that @xmath8 . in section [ sec : proof ]
we prove that this problem is np - complete .
this proof has implications on several variants of this problem ( section [ sec : corollaries ] ) .
in particular , no reasonable approximation algorithm exists , unless p@xmath9np . with the second approach we use the symmetric difference ( @xmath10 ) to quantify resemblance .
we now consider the full polygon rather than only its boundary , and look for a connected face set in @xmath2 instead of a cycle .
this leads to the following problem : [ prob : cfs ] let @xmath2 be a full grid graph , let @xmath1 be a simple polygon and let @xmath11 . decide whether a connected face set @xmath0 in @xmath2 exists such that @xmath12 . in spite of the symmetric difference being insensitive to matching different parts between polygons
, we prove that this is again an np - complete problem in section [ sec : cfs ] .
this result is independent of whether we allow holes in the resulting polygon .
moreover , the proof readily implies hardness for regular tilings using triangles or hexagons .
* schematization . * line and network schematization ( e.g. transit maps ) have received significant attention in the algorithmic literature , e.g. @xcite .
recently , schematization of geographic regions has gained increasing attention , e.g. @xcite .
our discretized approach is similar in nature to the octilinear schematization technique of cicerone and cermignani @xcite , though simplicity is of no concern in their work .
as mentioned in the introduction , the discretized approach offers conceptual advantages over the existing nondiscretized methods .
map matching has various applications in geographic information systems , such as finding a driven route based on a road network and a sequence of gps positions . without its simplicity constraint ,
it has been studied extensively with various criteria of resemblance , e.g. @xcite .
alt et al .
@xcite describe an algorithm that solves nonsimple map matching under the frchet distance in @xmath13 time where @xmath14 is the complexity of @xmath1 and @xmath15 the complexity of @xmath2 .
though `` u - turns '' can be avoided , no general simplicity guarantees are possible .
similarly , the decision problem for the weak frchet distance can be solved in @xmath16 time @xcite .
the _ simple _ map matching problem on the other hand has received little attention , although it stands to reason that for many applications a nonselfintersecting result is desired , if the input curve is simple .
a full grid graph always admits a solution with hausdorff distance at most @xmath17 and frchet distance at most @xmath18 , where @xmath19 parametrizes a realistic input model @xcite .
wylie and zhu @xcite prove independently that simple map matching under the _ discrete _ frchet distance is np - hard , however , without requiring a simple input curve .
a stronger result with a simple input curve follows directly from our proofs .
sherette and wenk @xcite show that it is np - hard to find a simple curve with bounded frchet distance on a 2d surface with holes or in 3d , but again without requiring a simple input curve . in its dual form ,
connected face selection is a specialization of the known np - hard maximum - weight - connected - subgraph problem @xcite in which a ( planar ) graph must be partitioned into two disjoint components , @xmath0 and @xmath20 , such that @xmath0 is connected and has maximal total weight .
our results readily imply that this dual problem remains np - hard even on a full grid graph if all weights are nonnegative and the size of @xmath0 is given ; this is independent of whether we constrain @xmath20 to be connected .
it is also related to the well - studied graph cuts , though focus there lies minimizing the number of cut edges connecting @xmath0 and @xmath20 ( e.g. @xcite ) .
vertex weights have been included only as part of the optimization criterion ( e.g. @xcite ) , the other part still being the number of cut edges .
the number of cut edges is not correlated to the complexity of the eventual shape : we can not use this as a trade - off between complexity and resemblance . also , these approaches tend not to require that a partition is connected , and focus on ensuring a certain balance between the two partition sizes . partitioning an unweighted nonplanar graph into connected components that each contain prescribed vertices is known to be np - hard @xcite
. * polygons . *
a polygon @xmath1 is defined by a cyclic sequence of vertices in @xmath21 .
each pair of consecutive vertices is connected by a line segment ( an edge ) .
a polygon is _ simple _ if no two edges intersect , except at common vertices .
we use @xmath22 to refer to the area of polygon @xmath1 .
the _ complexity _ of a polygon is its number of edges .
we use @xmath23 to refer to the boundary of @xmath1 .
unless mentioned otherwise , polygons are assumed to be simple throughout this paper .
a straight - line graph @xmath24 is defined by a set of vertices @xmath25 in @xmath21 and edges @xmath26 connecting pairs of vertices using line segments .
the graph is _ plane _ if no two edges intersect , except at common vertices . the _ complexity _ of a plane graph is its number of edges .
we call a plane graph a _ ( partial ) grid graph _ if all of its vertices have integer coordinates , and all edges are either horizontal or vertical , having length at least @xmath27
full _ grid graph is a maximal grid graph ( in terms of both vertices and edges ) within some rectangular region ; all edges have length @xmath27 .
a full grid graph represents a tiling of unit squares .
unless mentioned otherwise , graphs are assumed to be plane in this paper .
a _ cycle _ in a graph is a sequence of vertices such that every consecutive pair as well as the first and last vertex in the sequence are connected via an edge .
a cycle is _ simple _ if the sequence does not contain a vertex more than once .
a simple cycle in a plane graph corresponds to ( the boundary of ) a simple polygon .
the _ bends _ of a cycle are the vertices at which the interior angle is not equal to @xmath28 , that is , those that form corners in the polygon it represents .
complexity _ of a cycle is its number of bends .
the frchet distance quantifies the ( dis)similarity between two geometric shapes : a high frchet distance indicates a low similarity .
we define @xmath29 as the continuous function that maps the unit circle onto the boundary of @xmath1 .
let @xmath30 denote the set of all orientation - preserving homeomorphisms on @xmath31 . using @xmath32 to denote the euclidean distance , the frchet distance between two polygons is defined as @xmath33 maximal empty regions of a graph ( i.e. , not containing a vertex in its interior )
are referred to as _ faces_. one face , the outer face , is infinite ; all other faces have bounded area .
faces are said to be adjacent if they share at least one edge .
the dual @xmath34 of graph @xmath2 is defined by a vertex set containing a dual vertex for every face and two dual vertices are connected via a dual edge if the corresponding faces are adjacent .
a face set in @xmath2 is said to be _ connected _ if the corresponding induced subgraph of @xmath34 is connected .
we call a face set _ simply connected _ if the complement of the face set is also connected .
a simply connected face set corresponds to a simple polygon ; a ( nonsimply ) connected face set may have holes .
the symmetric difference between two polygons @xmath1 and @xmath35 is defined as the area covered by precisely one of the polygons .
that is , the symmetric difference is @xmath36 if @xmath37 is composed of pairwise disjoint regions @xmath38 ( e.g. a face set ) , then we may decompose the above formula into @xmath39 .
in this section we consider problem [ prob : smm ] , simple map matching .
we prove that this problem is np - complete , as formalized in the following theorem .
[ thm : main ] let @xmath2 be a partial grid graph , let @xmath1 be a simple polygon and let @xmath7 .
it is np - complete to decide whether @xmath2 contains a simple cycle @xmath3 with @xmath8 .
the problem is in np since the frchet distance can be computed in polynomial time @xcite and it is straightforward to check simplicity . in this section
we prove that the problem is also np - hard .
we conclude this section by considering the implications of this result on variants of this problem .
we assume @xmath40 in the remainder of this section .
de berg and khosravi @xcite prove that _ planar monotone 3sat _ is np - complete .
that is , decide whether a 3cnf formula @xmath41 is satisfiable , given the following constraints ( fig .
[ fig : planar3sat](a ) ) : * clauses are either positive ( only unnegated literals ) or negative ( only negated literals ) ; * a planar embedding for @xmath41 is given , representing variables and clauses as disjoint rectangles ; * variables lie on a single horizontal line ; * positive clauses lie above the variables , negative clauses below ; * links connecting clauses to the variables of their literals are strictly vertical . .
gray polygons represent gadgets , interact ingvia shared boundaries .
the red lines connect the various gadgets to obtain a simple polygon . ] for our reduction we construct a simple map matching instance a partial grid graph @xmath2 and a polygon @xmath1such that @xmath2 contains a simple cycle @xmath3 with @xmath42 if and only if formula @xmath41 is satisfiable .
we use three types of gadgets to represent the variables , clauses and links of @xmath41 : _ variable gadgets _ , _ clause gadgets _ , and
_ propagation gadgets_. our clause gadgets have small fixed dimensions and hence can not be stretched horizontally .
a single bend for the two `` outer '' edges of a clause is sufficient to ensure this ( see fig . [
fig : planar3sat](b ) ) . in the upcoming sections we first design the gadgets in isolation before completing the proof using the gadgets .
an overview of the eventual result of this construction is given in fig .
[ fig : constructionoverview ] .
each gadget specifies a _ local graph _ ( a part of @xmath2 ) and a _ local curve _ ( a part of @xmath1 ) .
the gadgets interact via vertices and edges shared by their local graphs .
there is no interaction based on the local curve : it is used only to force choices in using edges of the local graph .
if a cycle exists in the complete graph , a _ local path _ in the local graph must have a frchet distance of at most @xmath27 to the local curve .
the local path `` claims '' its vertices and edges : these can no longer be used by another gadget .
this results in _ pressure _ on the other gadget to use a different path .
a gadget has a number of pressure _ ports_. a port corresponds to a sequence of edges in the local graph that may be shared with another gadget .
a port may _ receive _ pressure , indicating that the shared edges and vertices may not be used in the gadget .
similarly , it may _ give _ pressure , indicating that the shared edges and vertices may not be used by any adjacent gadget .
all interaction between gadgets goes via these ports .
the local curves must be joined carefully to ensure that @xmath1 is simple . to this end
, each gadget has two curve _ gates _ that correspond to the endpoints of the local curve .
later , we show how to connect these gates to create a single simple polygon @xmath1 . in the following paragraphs
we describe the three gadgets .
in particular , we take note of the _ specification _ of each gadget : its behavior in terms of its ports ; a bounding polygon that contains the local graph and local curve ; the placement of its two gates and its ports .
these specifications capture all necessary aspects to complete the reduction in section [ ssec : proof ] . here
we focus on positive clauses ( i.e. , above the variables ) and their edges .
gadgets below are defined analogously , by mirroring vertically .
we present specifications and constructions mostly visually , using the following encoding scheme .
the bounding polygon is given with a black outline ; the ports are represented with thick green lines ; the gates are represented with red dots .
the local graph is given with thick light - blue lines ; the local curve is a red line .
we indicate various local paths using dark - blue lines ; ports that give pressure for a local path are indicated with an outward arrow .
all elements are visualized on an integer grid ( thin gray lines ) , to show that we indeed construct a partial grid graph .
all coordinates are an integer multiple of a half : all vertices are placed on vertices of the grid , exactly halfway an edge or exactly in the center of a cell . a clause gadget is illustrated in fig .
[ fig : clause_gadget ] .
it has fixed dimensions ; the figure precisely indicates its specification as well as its construction . the gadget admits a local path
only if one of its ports does _ not _ receive pressure .
any local path causes pressure on at least one port ; for each port there is a path that causes pressure only on that port .
the lack of external pressure on a port indicates that the value of the corresponding variable satisfies the clause . there is no local path that avoids all three ports :
if all ports receive pressure , none of the variables satisfies the clause and the gadget does not admit a local path .
the specification and construction of a variable gadget depend on the number of literals .
let @xmath43 denote the maximum of the number of positive and the number of negative literals of the variable .
we assume that @xmath44 : otherwise , it does not occur in any clause .
its bounding polygon , gates and ports are illustrated in fig .
[ fig : var_placement ] for @xmath45 . for higher values of @xmath43
, we increase the width by @xmath46 , to ensure a port of width @xmath47 and a distance of @xmath48 in between ports .
the gadget admits exactly two local paths : `` _ _ true _ _ '' and `` _ _ false _ _ '' .
ports for positive literals ( top side ) give pressure only with the _ false _ path .
ports for negative literals ( bottom side ) give pressure only with the _ true _ path .
in other words , a port gives pressure if the variable does _ not _ satisfy the corresponding clause . .
the right two figures indicate the two local paths , for @xmath49 . ] a propagation gadget ( shown in fig .
[ fig : prop_gadget ] ) connects a port of a variable gadget to a port of a clause gadget .
the bounding polygon is a corridor of width @xmath50 with at most one bend : if the link in formula @xmath41 has a bend , then the gadget also has a bend .
the corridor can have any integer height @xmath51 greater than @xmath52 .
if it has a bend , the corridor spans any integer width @xmath53 at least @xmath48 .
the two ports and gates of the gadget are placed as indicated .
the local graph and curve are constructed such that it admits only two local paths ; each puts pressure on exactly one port .
the gadget does not admit a path if both ports receive pressure .
if one port receives pressure , the other must give pressure : it propagates pressure .
we are now ready to construct graph @xmath2 and polygon @xmath1 based on formula @xmath41 . fig .
[ fig : constructionoverview ] illustrates this construction .
first , we place all variable gadgets next to one another , in the order determined by @xmath41 , with a distance of @xmath47 in between consecutive variables . using the @xmath54-coordinates in the embedding of @xmath41
, we sort the positive clauses to define a _ positive order _ @xmath55 .
we place the gadget for clause @xmath56 at a distance @xmath57 above the variables .
analogously , we use a _ negative order _ to place the negative clauses below the variables .
horizontally , the clause gadgets are placed such that the bottom port lines up with the appropriate port on the variable gadget of the middle literal . finally , we place a propagation gadget for each link in @xmath41 to connect the clause and variable gadgets . by placement of the clauses ,
any propagation gadget has height at least @xmath52 and a link without a bend can be represented by a propagation gadget without any bends .
as ports are at least distance @xmath48 apart , the width of a propagation gadget with a bend exceeds @xmath48 .
a propagation gadget does not overlap other gadgets : the placement of clauses would then imply that the provided embedding for @xmath41 is not planar . we have composed the various gadgets in polynomial time
however , we do not yet have a simple polygon .
we must `` stitch '' the local curves together ( in any order ) to create polygon @xmath1 . to this end
we first define three subcurves : @xmath58 for the variable gadgets ; @xmath59 for the positive clause gadgets and their propagation gadgets ; and @xmath60 for the negative clause gadgets and their propagation gadgets .
below is a detailed description of how these are constructed .
[ fig : constructionoverview ] visually illustrates the result . for @xmath58
we first define point @xmath61 and @xmath62 at distance @xmath50 outward from the leftmost and rightmost variable gadget respectively , both at the height of the gates .
we connect @xmath61 to the left gate of the leftmost variable , connect the matching gates of consecutive variables , and then connect the right gate of the rightmost variable to @xmath62 .
subcurve @xmath59 is constructed by defining a points @xmath63 and @xmath64 , similarly as for @xmath58 , but at the height of the `` positive '' propagation gadgets ; @xmath63 is placed at distance @xmath48 instead .
analogous to @xmath58 , we first create the straight traversal from @xmath63 to @xmath64 through all positive propagation gadgets .
we then include each positive clause in this subcurve , right before it `` enters '' the propagation gadget for its leftmost literal .
this is done by going up starting at distance @xmath65 before this gadget to two above the top side of the clause gadget , and then connecting to its right gate .
we go back from its left gate , and go down at distance @xmath27 before the propagation gadget .
now , @xmath59 traverses all positive clauses and their propagation gadgets .
subcurve @xmath60 is constructed analogous to @xmath59 , though @xmath63 is placed again at distance @xmath50 . by placement of the gadgets , these subcurves are simple polygonal curves between their respective endpoints ( @xmath66 and @xmath67 ) and do not intersect each other . to obtain a single simple polygon , we must now connect the endpoints of the three subcurves .
we connect @xmath62 to @xmath64 and @xmath61 to @xmath68 using vertical segments .
we connect @xmath69 to @xmath63 , by routing an edge at distance @xmath47 below the lowest negative clause .
we now have a simple polygon @xmath1 ; we define @xmath2 as the union of all local graphs and the parts of @xmath1 not contained in a gadget .
we now have constructed graph @xmath2 and polygon @xmath1 .
we must argue that the complexity is polynomial and that @xmath41 is satisfiable if and only if a simple cycle @xmath3 exists in @xmath2 with @xmath42 .
let @xmath15 denote the number of variables , and @xmath14 the number of clauses in @xmath41 .
the width of the construction is at most @xmath70 where @xmath71 is the number of occurrences of the @xmath72^th^ variable . as @xmath73 ,
the width of the construction is @xmath74 .
the height of the construction is at most @xmath75 , where @xmath76 is the number of positive clauses and @xmath77 the number of negative clauses ; since @xmath78 , the height is @xmath79 . as all coordinates are required to be an integer multiple of a half , this implies a polynomial bound on the complexity of @xmath2 and @xmath1 .
assume that @xmath41 is satisfiable and consider some satisfying assignment .
we argue the existence of a simple cycle @xmath3 . for each variable gadget
we choose either the `` true '' or `` false '' local path , matching the assigned truth value .
this gives pressure on a number of propagation gadgets : we choose the only remaining local path for these , causing pressure on the corresponding clauses .
for the other propagation gadgets , we choose the path such that it gives pressure at the variable and may receive pressure at the clause . since the truth values of the variables originate from a satisfying assignment , at most two ports of any clause receive pressure .
hence , the clause admits a local path as well .
we concatenate the local paths with the paths that are used to stitch together the local curves to obtain a simple cycle @xmath3 . by construction @xmath80
is at most @xmath27 . now , assume that @xmath2 contains a simple cycle @xmath3 with @xmath42 . by construction ,
cycle @xmath3 traverses all gadgets and contains exactly one local path for each gadget .
this local path ends at the gates of the gadget and the frchet distance between this local path and the local curve is at most @xmath27 . for a variable
, this local path corresponds to either the true or false state .
this directly yields the truth values of the variables .
each clause gadget also has a local path and hence one or more of its ports give pressure . since the propagation gadgets have a local path , the pressure from the clauses results in pressure on a variable gadget .
this pressure ensures that a variable that receives pressure from a clause is in a state satisfying the clause .
hence , the truth values found from the variables yield a satisfying assignment for formula @xmath41 .
this concludes the proof for theorem [ thm : main ] , showing that simple map matching is indeed np - complete .
this result and its construction have a number of implications , as discussed below . * approximation . * with the np - hardness result above , we may want to turn to approximation algorithms .
however , a simple argument shows that approximation is also np - hard .
[ cor : noapprox ] let @xmath2 be a partial grid graph of complexity @xmath14 and let @xmath1 be a simple polygon of complexity @xmath15 .
it is np - hard to approximate the minimal @xmath80 of any simple cycle @xmath3 in @xmath2 within any factor of @xmath81 where @xmath82 is a polynomial in @xmath15 and @xmath14 . for an unsatisfiable formula
the minimal frchet distance of a simple cycle in the constructed graph is significantly larger than @xmath27 .
suppose that this minimal frchet distance is strictly greater than @xmath83 .
any @xmath83-approximation algorithm for simple map matching is able to decide satisfiability of planar monotone 3sat formulas .
thus , unless p@xmath9np , no @xmath83-approximation algorithm can have polynomial execution time . to determine the exact value of @xmath83
, we wish to determine at which value a local path becomes admissible that does not correspond to the desired behavior of the construction .
this occurs at @xmath84 : the clause gadget breaks , admitting a local path that does not pass through any of the ports .
however , it is straightforward to lengthen the gadgets to increase the value of @xmath83 . this does not increase the combinatorial complexity of @xmath2 and @xmath1 , thus maintaining a polynomial - size instance .
the construction now spans an area @xmath85 by @xmath86 , where @xmath87 and @xmath88 are the number of variables and clauses in formula @xmath41 .
thus , we require @xmath89 to encode the coordinates in polynomial space .
r0.3 * counting bends . * for schematization
we do not wish to just find some cycle in @xmath2 , but also optimize or bound its complexity , measured in the number of bends .
if @xmath2 is a partial grid graph ( corresponding to rectilinear schematization ) , any bend is either a left or right turn .
every rectilinear polygon has a certain _ bend profile _ :
the sequence of left and right bends in counterclockwise order along its boundary .
the bend profile gives no information about edge lengths : seemingly different polygons have the same bend profile ( see fig . [
fig : turnprofile ] ) . unfortunately , even with a given bend profile , the problem remains np - complete .
however , in this reduction the bend profile has length proportional to the complexity of the formula .
it does not prove that no fixed - parameter - tractable ( fpt ) algorithm exists .
[ cor : nobends ] let @xmath2 be a partial grid graph , let @xmath1 be a simple polygon , let @xmath7 and let @xmath90 be a bend profile .
it is np - complete to decide whether @xmath2 contains a simple cycle @xmath3 that adheres to @xmath90 with @xmath8 . in all constructions , the bends made by the various local paths are identical .
we can easily derive a bend profile that must lead to a simple cycle , if the formula is satisfiable .
* area preservation .
* suppose we want an area - preserving solution , i.e. , the area of cycle @xmath3 must be equal to that of @xmath1 .
a simple argument proves that our reduction can be extended to prove that this more constrained problem is also np - complete .
[ cor : smm - area ] let @xmath2 be a partial grid graph , let @xmath1 be a simple polygon and let @xmath7 .
it is np - complete to decide whether @xmath2 contains a simple cycle @xmath3 with @xmath8 and @xmath91 .
[ fig : constructionoverview_complement ] at the end of the paper shows an overview of the modified construction .
the key observation is that the original construction is duplicated , but inside and outside the polygon have been inverted .
this is achieved by connecting the endpoints of the subcurves @xmath92 , @xmath93 and @xmath94 , slightly differently , i.e. , with those of the duplicate .
any solution @xmath3 coincides with @xmath1 outside the gadgets : only the local paths change the area of @xmath3 .
hence , any change in area resulting from a local path in one of the gadgets in the one copy can be counteracted by choosing the exact same local path in the other copy .
* variants . *
finally , there are a number of variants of the problem that can be proven to be np - complete via the same construction . as strict monotonicity in the homeomorphism
is not crucial for the reduction , the problem under the _ weak _ frchet distance is also np - complete .
the clause gadget admits an extra local path , one that still exhibits the desired behavior .
the problem is also np - complete under the _ discrete _ ( weak ) frchet distance , as we may sample the graph and polygon appropriately .
not every grid location needs a vertex , preserving the inapproximability result of corollary [ cor : noapprox ] . as all interaction between gadgets
is based on edges , it is also np - complete to determine the existence of an `` edge - simple '' cycle that uses each edge at most once ( but vertices may be used more than once ) .
it is not essential in any of the above ( except for area preservation ) for @xmath1 to be a closed curve : the same construction works for open curves , and thus these variants are np - hard as well .
with the negative results using the frchet distance , we now consider the symmetric difference via problem [ prob : cfs ] , connected face selection .
we prove that this problem is also np - complete , even on a full grid graph , as captured in the following theorem .
[ thm : cfs ] let @xmath2 be a full grid graph , let @xmath1 be a simple polygon and let @xmath11 .
it is np - complete to decide whether @xmath2 contains a connected face set @xmath0 with @xmath12 .
the problem is obviously in np , since the symmetric difference and connectedness can be straightforwardly verified in polynomial time .
here we show that it is also np - hard .
we conclude this section with some implications of this result
. * rectilinear steiner tree . *
the rectilinear steiner tree problem is formulated as follows : given a set @xmath95 of @xmath15 points in @xmath21 , is there a tree @xmath90 of total edge length at most @xmath96 that connects all points in @xmath95 , using only horizontal and vertical line segments ? vertices of @xmath90 are not restricted to @xmath95 .
this problem was proven np - complete by garey and johnson @xcite .
hanan @xcite showed that an optimal result must be contained in graph @xmath97 corresponding to the arrangement of horizontal and vertical lines through each point in @xmath95 ( see fig . [
fig : cfs_sketch](a ) ) .
subsequently , this was called the _ hanan grid _ ( e.g. @xcite ) .
we call a vertex of @xmath97 a _ node _ if it corresponds to a point in @xmath95 and a _ junction _ otherwise .
as the problem is scale invariant , we assume @xmath98 .
all edges in @xmath97 must be shorter than @xmath27 : otherwise , the answer is trivial no such tree exists .
( dark points ) .
( b ) @xmath2 consists of unit square cells : node - cells are colored black , junction - cells white , edge - cells gray ; face - cells are hatched .
( c ) `` skeleton '' @xmath99 for polygon @xmath1 .
( d ) sketch of polygon @xmath1 constructed on top @xmath2 . ; light blue areas indicate flexible parts in the construction that represent edge lengths . ]
we must transform point set @xmath95 into a full grid graph @xmath2 , a polygon @xmath1 and a value @xmath11 .
we construct @xmath2 such that each _ cell _ ( face of @xmath2 ) corresponds to a vertex ( _ node - cell _ or _ junction - cell _ ) , an edge ( _ edge - cell _ ) , or a bounded face ( _ face - cell _ ) of @xmath97 ; see fig .
[ fig : cfs_sketch](b ) .
we then construct polygon @xmath1 by defining a part of the polygon inside all cells , except for the face - cells : @xmath1 does not overlap these . to structure @xmath1
we use a _ skeleton _
@xmath99 , a tree spanning the non - face - cells in the dual of @xmath2 ( fig .
[ fig : cfs_sketch](c ) ) .
recall that the symmetric difference between @xmath1 and a face set @xmath100 may be computed as @xmath101 .
as @xmath102 , we define the _ weight _ of a cell @xmath83 in @xmath2 as @xmath103 . hence , the symmetric difference is @xmath104 .
we set the desired weight @xmath105 for cell @xmath83 to : * if @xmath83 is a node - cell ; * if @xmath83 is a junction - cell ; * if @xmath83 is an edge - cell , where @xmath106 is the length of the corresponding edge @xmath107 in @xmath97 ; * if @xmath83 is a face - cell .
given a desired weight @xmath105 for cell @xmath83 , the area of overlap @xmath108 is @xmath109 .
every cell has @xmath108 area of @xmath1 inside ; @xmath22 equals @xmath110 .
we call @xmath111 the _ local polygon _ of @xmath83 .
we set @xmath112 .
that is , the sum of weights is at most @xmath113 .
this can be achieved only if the face set contains all node - cells and no face - cell .
we design every cell such that the desired weight is achieved . for a face - cell , this is trivial : @xmath114 , hence @xmath115 and we keep @xmath1 disjoint from this cell . for all other cells @xmath1 to cover some fraction of its interior , as dictated by @xmath108 .
skeleton @xmath99 dictates how to connect the local polygons ; we ensure that at least the middle @xmath116 of the shared edge ( the _ connector _ ) is covered .
a local polygon should never touch the corners of its cell .
node- and junction - cells may have up to four neighbors in @xmath99 . covering the four connectors is done with a cross shape , covering @xmath117th of the cell s area . by thickening this cross
, we can straightforwardly support @xmath108 with @xmath118 , ensuring not to touch the corners or edges shared with cells that are not adjacent in @xmath99 .
a node - cell has weight @xmath119 ; @xmath120 .
a junction - cell has weight @xmath121 ; @xmath122 . hence ,
both can be represented ( see fig . [
fig : cfs_design](a b ) ) .
an edge - cell has weight @xmath123 and thus should cope with weights between zero and a half ; @xmath108 lies between @xmath124 and @xmath125 .
any edge - cell has degree @xmath27 or @xmath47 in @xmath99 ; if it has degree @xmath47 , the neighboring cells are on opposite sides .
hence , @xmath126 can be trivially handled by creating a rectangular shape that touches exactly the necessary connectors .
@xmath122 is dealt with by widening this rectangle within the cell ; any intermediate weight is handled by interpolating between these two .
this is illustrated in fig .
[ fig : cfs_design](c ) . by @xmath1 .
( b ) junction - cells are covered for @xmath127 by @xmath1 .
( c ) between @xmath116 ( dark blue ) and @xmath127 ( dark and light blue ) of edge - cells are covered by @xmath1 . ]
we now have graph @xmath2 , polygon @xmath1 and a value for @xmath128 .
we must prove that the reduction is polynomial and that @xmath95 has a rectilinear steiner tree of length at most @xmath27 , if and only if there is a connected face set @xmath0 in @xmath2 such that @xmath12 .
the former is trivial by observing that : @xmath1 has @xmath129 complexity in each of the @xmath130 cells ; computing tree @xmath90 for guiding the construction of @xmath1 can be done with for example a simple breadth - first search .
suppose we have a rectilinear steiner tree @xmath90 of length at most @xmath27 in the hanan grid .
we construct a face set @xmath0 as the union of all cells corresponding to vertices and edges in @xmath90 . by definition of @xmath90 , this must contain all node - cells and can not contain face - cells . as junction - cells
have no weight , the total weight of @xmath0 is @xmath131 where @xmath132 is the cell of @xmath2 corresponding to edge @xmath107 . by assumption @xmath133 : the total weight is at most @xmath113 .
thus , the symmetric difference for @xmath0 is at most @xmath134 .
suppose we have a connected face set @xmath0 in @xmath2 such that @xmath12 .
the total weight is thus @xmath135 .
since face - cells have weight @xmath27 and only node - cells have negative weight , being @xmath119 , this can be achieved only if @xmath0 contains all node - cells and no face - cells .
in particular , the sum of the weights over all edge - cells is at most @xmath125 .
thus , the subgraph of @xmath97 described by the selected cells must be connected , contain all nodes of @xmath95 , and have total length at most @xmath27 .
if this subgraph is not a tree , we can make it a tree , by leaving out edges ( further reducing the total length ) , until the subgraph is a tree . *
simply connected .
* the same reduction works for a _ simply _ connected face set , as steiner tree @xmath90 can not contain cycles and a simply connected face set in @xmath2 readily implies a tree .
let @xmath2 be a full grid graph , let @xmath1 be a simple polygon and let @xmath11 .
it is np - complete to decide whether @xmath2 contains a simply connected face set @xmath0 with @xmath12 .
the face set obtained from a steiner tree must be simply connected , since @xmath90 can not contain cycles .
in addition , a simply connected face set in @xmath2 ( that does not contain a face - cell ) directly describes a tree
. * area preservation . * with an area - preservation constraint the problem remains np - complete .
we only sketch an argument ; full proof can be found in appendix [ app : cfs - impl ] .
this readily implies that variants prescribing the number of faces or total area via a parameter are also np - hard .
[ cor : cfs - area ] let @xmath2 be a full grid graph , let @xmath1 be a simple polygon and let @xmath11 .
it is np - complete to decide whether @xmath2 contains a ( simply ) connected face set @xmath0 with @xmath12 and @xmath136 .
we assume for simplicity that no points in @xmath95 share an @xmath137- or @xmath54-coordinate : @xmath97 is then exactly an @xmath138 grid .
the number of cells we need to represent a steiner tree spanning @xmath95 is at least @xmath139 and at most @xmath140 .
thus , we need @xmath141 , as to allow sufficiently many cells to be selected in the original construction . to achieve the necessary area for @xmath1 , we are going to add more cells to the construction , each with @xmath122 and thus @xmath142 .
we add @xmath143 such cells . due to their weight
, their presence in or absence from face set @xmath0 has no effect on the symmetric difference .
we need the new cells to not interfere with the original construction .
therefore , they should be adjacent only with one of the node - cells on the boundary of the construction .
thus we also add @xmath144 cells weight @xmath114 to separate the new cells from the original cells .
[ fig : cfs_sketcharea ] illustrates an overview of this extended construction and its new skeleton .
note that the newly added cells have the same weights as the original junction - cells and face - cells ; the construction of @xmath1 thus extends straightforwardly .
if we are given some connected face set ( that now uses a fixed number of cells ) that has sufficiently low symmetric difference , we still know that it must span all the node - cells and thus encode a rectilinear steiner tree of sufficiently low length . to transform a rectilinear steiner tree @xmath90 to a connected face set , we must argue that we can always select the correct number of cells .
first , let us bound the size of @xmath22 .
since edges have @xmath108 in between @xmath124 and @xmath125 , it is easy to derive that the total area @xmath145 of the original construction satisfies @xmath146 ; this simplifies to @xmath147 .
the new cells add @xmath148 area .
thus , @xmath149 is bounded by the interval @xmath150 $ ] . hence , the number of cells we need to select is more than @xmath151 and strictly less than the number of newly added cells with @xmath142
. therefore , we simply apply the original transformation , but use the new cells as `` overflow '' for any cells in excess of those needed to represent @xmath90 .
the above assumes that the eventual area of @xmath1 is an integer ; since it depends on the edge lengths in @xmath97 , it likely is not .
this can be remedied by either assuming we are allowed to round the area of @xmath1 , or by letting @xmath1 extend slightly into the outer face to make it integer .
note that the weight of the outer face is always infinite for a bounded @xmath1 .
alternatively , we can use any weight for node - cells , strictly in between @xmath125 and @xmath27 .
more general graph classes ( e.g. plane graphs ) are also np - hard ; this is readily implied by theorem [ thm : cfs ] .
finally , the problem remains np - complete for graphs representing hexagonal and triangular tilings ( by combining two triangles into one slanted square ) .
schematic maps are an important tool to visualize data with a geographic component .
we studied discretized approaches to their construction , by restricting solutions to a plane graph @xmath2 .
this promotes alignment and uniformity of edge lengths and avoids the risk of visual collapse .
we considered two variants : _ simple map matching _ using the frchet distance and _ connected face selection _ using the symmetric difference .
unfortunately , both turn out to be np - complete ; the former is even np - hard to approximate .
the proofs readily imply that a number of variants are np - hard as well , even with an area - preservation constraint .
it remains open whether ( general ) simple map matching is fixed - parameter tractable .
moreover , there is quite a gap between the graph necessary for the reduction and the graphs that would be useful in the context of schematization ( e.g. full grid graphs , triangular tilings ) .
are such instances solvable in polynomial time ? for connected face selection , we know hardness on such constructed graphs , but it remains open whether approximation or fpt algorithms are possible . in both methods the reduction needs rather convoluted polygons , very unlike the geographic regions that we want to schematize : do realistic input assumptions help to obtain efficient algorithms ?
recently , bouts et al .
@xcite have show that any full grid graph admits a simple cycle with frchet distance bounded using a realistic input model , called narrowness .
however , this does not readily preclude the decision problem from being np - hard , even for polygons that have bounded narrowness .
the frchet distance is a bottleneck measure and thus results obtained via simple map matching may locally deviate more than necessary , even when minimizing the number of bends . buchin et al .
@xcite introduced `` locally correct frchet matchings '' to counteract this flaw with the frchet distance .
can we extend this concept to simple map matching ?
the author would like to thank : bettina speckmann and her applied geometric algorithms group , kevin buchin , bart jansen , arthur van goethem , marc van kreveld and aidan slingsby for inspiring discussions on the topic of this paper ; gerhard woeginger for pointing out the exponential bound in corollary [ cor : noapprox ] .
the author was partially supported by the netherlands organisation for scientific research project 639.023.208 and marie skodowska - curie action msca - h2020-if-2014 656741 . c. wenk ,
r. salas & d. pfoser . addressing the need for map - matching speed : localizing global curve - matching algorithms . in _ proc .
conf . on sci . & stat .
database management _ ,
pages 379388 , 2006 . | to produce cartographic maps , simplification is typically used to reduce complexity of the map to a legible level . with schematic maps ,
however , this simplification is pushed far beyond the legibility threshold and is instead constrained by functional need and resemblance . moreover ,
stylistic geometry is often used to convey the schematic nature of the map . in this paper
we explore discretized approaches to computing a schematic shape @xmath0 for a simple polygon @xmath1 .
we do so by overlaying a plane graph @xmath2 on @xmath1 as the solution space for the schematic shape .
topological constraints imply that @xmath0 should describe a simple polygon .
we investigate two approaches , _ simple map matching _ and _ connected face selection _ , based on commonly used similarity metrics . with the former , @xmath0 is a simple cycle @xmath3 in @xmath2 and we quantify resemblance via the frchet distance .
we prove that it is np - hard to compute a cycle that approximates the minimal frchet distance over all simple cycles in a plane graph @xmath2 .
this result holds even if @xmath2 is a partial grid graph , if area preservation is required and if we assume a given sequence of turns is specified . with the latter
, @xmath0 is a connected face set in @xmath2 , quantifying resemblance via the symmetric difference .
though the symmetric difference seems a less strict measure , we prove that it is np - hard to compute the optimal face set .
this result holds even if @xmath2 is full grid graph or a triangular or hexagonal tiling , and if area preservation is required .
moreover , it is independent of whether we allow the set of faces to have holes or not . |
in recent years there has been growing interst in graph polynomials , functions from graphs to polynomial rings which are invariant under isomorphism .
graph polynomials encode information about the graphs in a compact way in their evaulations , coeffcients , degree and roots .
therefore , efficient computation of graph polynomials has received considerable attention in the literature . since most graph polynomials which naturally arise are @xmath6-hard to compute ( see e.g. @xcite ) , a natural perspective under which to study the complexity of graph polynomials
is that of _
parameterized complexity_. parameterized complexity is a successful approach to tackling @xmath7-hard problems @xcite , by measuring complexity with respect to an additional _ parameter _ of the input ; we will be interested in the parameters tree - width and clique - width .
a computational problem is _ * fixed - parameter tractable * _ ( @xmath1 ) with respect to a parameter @xmath5 if it can be solved in time @xmath8 , where @xmath9 is a computable function of @xmath5 , @xmath4 is the size of the input , and @xmath10 is a polynomial in @xmath4 .
many @xmath7-hard problems are fixed parameter tractable for an appropriate choice of parameter , see @xcite for many examples . every problem in the infinite class of decision problems definable in _ monadic second order logic _ ( @xmath0 ) is fixed - parameter tractable with respect to tree - width by courcelle s theorem @xcite ( though the result originally was not phrased in terms of parameterized complexity ) .
the computation problem we consider for a graph polynomial @xmath11 is the following : @xmath12 @xmath13 : a graph @xmath14 @xmath15 compute the coefficients @xmath16 of the monomials @xmath17 . for graph polynomials , a parameterized complexity theory with respect to tree - width has been developed . here
, the goal is to compute , given an input graph , the table of coefficients of the graph polynomial .
the tutte polynomial has been shown to be fixed - parameter tractable @xcite .
@xcite used a logical method to study the parameterized complexity of an infinite class of graph polynomials , including the tutte polynomial , the matching polynomial , the independence polynomial and the ising polynomial .
@xcite showed that the class of graph polynomials definable in @xmath0 in the vocabulary of hypergraphs in the vocabulary of hypergraphs is denoted @xmath18 , while @xmath0 in the vocabulary of graphs is denoted @xmath19 . ]
is fixed - parameter tractable .
this class contains the vast majority of graph polynomials which are of interest in the literature . going beyond tree - width to clique -
width the situation becomes more complicated .
@xcite studied the class of graph polynomials definable in @xmath0 _ in the vocabulary of graphs_. they proved that every graph polynomial in this class is fixed - parameter tractable with respect to clique - width .
however , this class of graph polynomials does not contain important examples such as the chromatic polynomial , the tutte polynomial and the matching polynomial .
in fact , @xcite proved that the chromatic polynomial and the tutte polynomial are not fixed - parameter tractable with respect to clique - width ( under the widely believed complexity - theoretic assumption that @xmath20 $ ] ) .
@xcite proved that the chromatic polynomial and the matching polynomial are _ * fixed - parameter polynomial time * _ computable with respect to clique - width , meaning that they can be computed in time @xmath21 , where @xmath4 is the size of the graph , @xmath22 is the clique - width of the graph and @xmath9 is a computable function . @xcite proved an analogous result for the ising polynomial .
the main result of this paper is a meta - theorem generalizing the fixed - parameter polynomial time computability of the chromatic polynomial , the matching polynomial and the ising polynomial to an infinite family of graph polynomials definable in @xmath0 analogous to @xcite .
let @xmath23 be an @xmath0-ising polynomial .
@xmath23 is fixed - parameter polynomial time computable with respect to clique - width .
the class of @xmath0-ising polynomials is defined in section [ se : msolising ] .
let @xmath24=\left\ { 1,\ldots , k\right\}$ ] .
let @xmath25 be the vocabulary of graphs @xmath26 consisting of a single binary relation symbol @xmath27 .
a @xmath5-graph is a structure @xmath28 which consists of a simple graph @xmath29 together with a partition @xmath30 of @xmath31 .
let @xmath32 denote the vocabulary of @xmath5-graphs @xmath33 extending @xmath25 with unary relation symbols @xmath34 .
the class @xmath35 of @xmath5-graphs of clique - width at most @xmath5 is defined inductively .
singletons belong to @xmath35 , and @xmath35 is closed under disjoint union @xmath36 and two other operations , @xmath37 and @xmath38 , to be defined next . for any @xmath39 $ ] , @xmath40 is obtained by relabeling any vertex with label @xmath41 to label @xmath42 . for any @xmath39 $ ] ,
@xmath43 is obtained by adding all possible edges @xmath44 between members of @xmath41 and members of @xmath42 .
the clique - width of a graph @xmath14 is the minimal @xmath5 such that there exists a labeling @xmath45 for which @xmath46 belongs to @xmath35 .
we denote the clique - width of @xmath14 by @xmath22 .
the clique - width operations @xmath37 and @xmath38 are well - defined for @xmath5-graphs .
the definitions of these operations extend naturally to structures @xmath47 which expand @xmath5-graphs with @xmath48 .
a @xmath5-expression is a term @xmath49 which consists of singletons , disjoint unions @xmath36 , relabeling @xmath37 and edge creations @xmath38 , which witnesses that the graph @xmath50 obtained by performing the operations on the singletons is of clique - width at most @xmath5 .
every graph of tree - width at most @xmath5 is of clique - width at most @xmath51 , cf . @xcite .
while computing the clique - width of a graph is @xmath7-hard , s. oum and p. seymour showed that given a graph of clique - width @xmath5 , finding a @xmath52-expression is fixed parameter tractable with clique - width as parameter , cf .
@xcite . for
a formula @xmath53 , let @xmath54 denote the quantifier rank of @xmath53 .
for every @xmath55 and vocabulary @xmath56 , we denote by @xmath57 the set of @xmath0-formulas on the vocabulary @xmath56 which have quantifier rank at most @xmath58 . for two @xmath56-structures @xmath59 and @xmath60 , we write @xmath61 to denote that @xmath59 and @xmath60 agree on all the sentences of quantifier rank @xmath58 . an @xmath62-ary operation @xmath63 on @xmath56-structures is called _ smooth _
if for all @xmath64 , whenever @xmath65 for all @xmath66 , we have @xmath67 smoothness of the clique - width operations is an important technical tool for us : [ smoothness , cf .
@xcite ] [ th : smooth ] 1 . for every vocabulary @xmath56 ,
the disjoint union @xmath36 of two @xmath56-structures is smooth .
2 . for every @xmath68 , @xmath37 and @xmath38
are smooth .
it is convenient to reformulate theorem [ th : smooth ] in terms of _ hintikka sentences _ ( see @xcite ) : [ hintikka sentences ] [ prop : hin ]
let @xmath56 be a vocabulary .
for every @xmath69 there is a finite set @xmath70 of @xmath57-sentences such that 1 .
every @xmath71 has a finite model .
2 . the conjunction @xmath72 of any two distinct @xmath73 is unsatisfiable .
every @xmath57-sentence @xmath74 is equivalent to exactly one finite disjunction of sentences in @xmath75 .
4 . every @xmath56-structure @xmath59 satisfies a unique member @xmath76 of @xmath75 . in order to simplify notation
we omit the subscript @xmath56 in @xmath77 when @xmath56 is clear from the context .
let @xmath78 the be the vocabulary consisting of the binary relation symbol @xmath79 and the unary relation symbol @xmath80 .
let @xmath81 extend @xmath78 with the unary relation symbols @xmath82 . from theorem [ th : smooth ] and proposition [ prop : hin ] we get : for every @xmath83 : 1 .
there is @xmath84 such that , for every @xmath85 and @xmath86 , @xmath87 .
2 . for every unary operation @xmath88\}$ ] , there is @xmath89 such that , for every @xmath90 , @xmath91 . for every @xmath92 ,
let @xmath93 , where @xmath94 are new unary relation symbols .
[ @xmath0-ising polynomials ] [ def : msolising ] for every @xmath95 , @xmath96 and @xmath29 we define @xmath97 as follows : @xmath98 @xmath99 is the sum over partitions @xmath100 of @xmath31 such that @xmath101 satisfies @xmath74 of the monomials obtained as the product of @xmath102 for all @xmath103 and @xmath104 for all @xmath105 .
[ ising polynomial ] the trivariate ising polynomial @xmath106 is a partition function of the ising model from statistical mechanics used to study phase transitions in physical systems in the case of constant energies and external field .
@xmath106 is given by @xmath107 where @xmath108 denotes the set of edges between @xmath109 and @xmath110 , and @xmath111 denotes the set of edges inside @xmath109 .
@xmath106 was the focus of study in terms of hardness of approximation in @xcite and in terms of hardness of computation under the exponential time hypothesis was studied in @xcite .
@xcite also showed that @xmath106 is fixed - parameter polynomial time computable .
@xmath112 generalizes a bivariate ising polynomial , which was studied for its combinatorial properties in @xcite .
@xcite showed that @xmath106 contains the matching polynomial , the van der waerden polynomial , the cut polynomial , and , on regular graphs , the independence polynomial and clique polynomial .
the evaluation of @xmath113 at @xmath114 , @xmath115 , @xmath116 , @xmath117 and @xmath118 gives @xmath106 and therefore @xmath106 is an @xmath0-ising polynomial .
[ independence - ising polynomial ] the independence - ising polynomial @xmath119 is given by @xmath120 @xmath119 contains the independence polynomial as the evaluation @xmath121 .
see the survey @xcite for a bibliography on the independence polynomial .
the evaluation @xmath122 is @xmath123 , where @xmath124 is the number of isolated vertices in @xmath14 .
@xmath119 is an evaluation of an @xmath0-ising polynomial : @xmath125 where @xmath126 .
[ dominating - ising polynomial ] the dominating - ising polynomial is given by @xmath127 @xmath128 where @xmath108 denotes the set of edges between @xmath109 and @xmath110 .
@xmath127 contains the domination polynomial @xmath129 .
@xmath129 is the generating function of its dominating sets and we have @xmath130 .
the domination polynomial first studied in @xcite and it and its variations have received considerable attention in the literature in the last few years , see e.g. @xcite .
previous research focused on combinatorial properties such as recurrence relations and location of roots .
hardness of computation was addressed in @xcite .
@xmath127 encodes the degrees of the vertices of @xmath14 : the number of vertices with degree @xmath131 is the coefficient of @xmath132 in @xmath127 .
@xmath127 is an @xmath0-ising polynomial given by @xmath133 , where @xmath134 two classes of graph polynomials which have received attention in the literature are : 1 .
@xmath0-polynomials on the vocabulary of graphs , and 2 .
@xmath0-polynomials on the vocabulary of hypergraphs .
see e.g. @xcite for the exact definitions .
the former class contains graph polynomials such as the independence polynomial and the domination polynomial .
the latter class contains graph polynomials such as the tutte polynomial and the matching polynomial .
every graph polynomial which is @xmath0-definable on the vocabulary of graphs is also @xmath0-definable on the vocabulary of hypergraphs .
the class of @xmath0-ising polynomials strictly contains the @xmath0-polynomials on graphs , see fig .
the containment is by definition . for the strictness
, we use the fact that by definition the maximal degree of any indeterminate in an @xmath0-polynomial on graphs grows at most linearly in the number of vertices , while the maximal degree of @xmath135 in the ising polynomial @xmath136 of the complete bipartite graph @xmath137 equals @xmath138 .
every @xmath0-ising polynomial @xmath99 is an @xmath0-polynomial on the vocabulary of hypergraphs , given e.g. by @xmath139 where the summation over @xmath140 is exactly as in definition [ def : msolising ] , and the summation over @xmath141 is over tuples @xmath142 of subsets of the edge set of @xmath14 satisfying @xmath143 , where @xmath144 we use the fact that @xmath145 is a partition of the set of vertices is definable in @xmath0 .
+ containments of classes of graph polynomials definable in @xmath0 . ,
title="fig : " ]
we are now ready to state the main theorem and prove a representative case of it .
[ main theorem][th : main ] for every @xmath0-ising polynomial @xmath99 there is a function @xmath146 such that @xmath147 is computable on graphs @xmath14 of size @xmath4 and of clique - width at most @xmath5 in running time @xmath148 .
we prove the theorem for graph polynomials of the form @xmath149 for every @xmath150 .
the summation in @xmath151 is over subsets @xmath109 of the vertex set of @xmath14 .
the graph polynomials @xmath151 are a notational variation of @xmath99 with @xmath152 , @xmath115 and @xmath153 : for every @xmath154 , @xmath155 , where @xmath156 is obtained from @xmath74 by substituting @xmath157 with @xmath158 and @xmath159 with @xmath160 .
the proof for the general case is in similar spirit . for every @xmath55
there is a finite set @xmath161 of @xmath162-ising polynomials such that , for every formula @xmath163 , @xmath151 is a sum of members of @xmath161 ( see below ) .
the algorithm computes the values of the members of @xmath161 on @xmath14 by dynamic programming over the parse term of @xmath14 , and using those values , the value of @xmath164 on @xmath14 .
more precisely , for every @xmath165 , let @xmath166 and let @xmath167 every @xmath168 also belongs to @xmath169 , and hence there exists by proposition [ prop : hin ] a set @xmath170 such that @xmath171 hence , @xmath172 setting @xmath173 and @xmath174 for all @xmath175 . for tuples @xmath176),(b_{c_{1},c_{2}}:c_{1},c_{2}\in[k])\right)\in[n]^{k}\times[n]^{k^2}$ ] , let @xmath177 be the coefficient of @xmath178 in @xmath179
. given a @xmath5-graph @xmath14 , the algorithm first computes a parse tree @xmath49 as in @xcite .
the algorithm then computes @xmath179 for all @xmath165 by induction over @xmath49 : 1 .
if @xmath14 is a graph of size @xmath180 , then @xmath181 is computed directly .
2 . let @xmath14 be the disjoint union of @xmath182 and @xmath183 .
we compute @xmath184 for every @xmath185 and @xmath186^{k}\times[n]^{k^2}$ ] as follows :
@xmath187 3 . let @xmath188 .
we compute @xmath184 for every @xmath165 and @xmath186^{k}\times[n]^{k^2}$ ] as follows : @xmath189 where the inner summation is over @xmath190 such that @xmath191 and @xmath192 4 .
let @xmath193 with @xmath194 .
let @xmath195 be the number of vertices in @xmath14 .
we compute @xmath184 for every @xmath165 and @xmath186^{k}\times[n]^{k^2}$ ] as follows : @xmath196 where the summation is over @xmath190 such that @xmath197 and @xmath198 finally , the algorithm computes @xmath151 as the sum from eq .
( [ eq : sumqa ] ) .
the main observations for the runtime analysis are : * the size of the set @xmath199 of hintikka sentences is a function of @xmath5 but does not depend on @xmath4 .
let @xmath200 .
* by definition of @xmath201 , for a monomial @xmath202 to have a non - zero coefficient , it must hold that @xmath203 and @xmath204 , since @xmath205 and @xmath206 are sizes of sets of vertices and sets of edges , respectively . *
the coefficient of any monomial of @xmath201 is at most @xmath207 . *
the parse tree guaranteed in @xcite is of size @xmath208 for suitable @xmath209 and @xmath210 .
the algorithm performs a single operation for every node of the parse tree .
* singletons * : the coefficients of every @xmath211 for a singleton @xmath5-graph can be computed in time @xmath212 , which can be bounded by @xmath213 .
* disjoint union * , * recoloring * * and * * edge additions * : the algorithm sums over ( 1 ) @xmath214 or pairs @xmath215 and ( 2 ) over @xmath216^{k}\times[n]^{k^2}$ ] or pairs @xmath217^{k}\times[n]^{k^2}\right)^{2}$ ] , then ( 3 ) performs a fixed number of arithmetic operations on numbers which can be written in @xmath218 space .
each node in the parse tree requires time at most @xmath219^{k}\times[n]^{k^2}\right)^{2}\right)$ ] .
since the size of the parse tree is @xmath208 , the algorithm runs in fixed - parameter polynomial time .
we have defined a new class of graph polynomials , the @xmath0-ising polynomials , extending the @xmath0-polynomials on the vocabulary of graphs and have shown that every @xmath0-ising polynomial can be computed in fixed - parameter polynomial time .
this result raises the question of which graph polynomials are @xmath0-ising polynomials . in previous work @xcite we have developed a method based on connection matrices to show that graph polynomials are not definable in @xmath0 over either the vocabulary of graphs or hypergraphs .
the tutte polynomial does not seem to be an @xmath0-ising polynomial .
@xcite proved that the tutte polynomial can be computed in subexponential time for graphs of bounded clique - width .
more precisely , the time bound in @xcite is of the form @xmath220 , where @xmath221 for all @xmath222 .
is there a natural infinite class of graph polynomials definable in @xmath0 which includes the tutte polynomial such that membership in this class implies _ fixed parameter subexponential time _
computability with respect to clique - width ( i.e. , that the graph polynomial is computable in @xmath223 time for some function @xmath224 satisfiying @xmath225 for all @xmath222 ) ?
fedor v. fomin , petr a. golovach , daniel lokshtanov , and saket saurabh .
algorithmic lower bounds for problems parameterized with clique - width . in _ proceedings of the twenty - first annual acm - siam symposium on discrete algorithms , soda 2010 ,
austin , texas , usa , january 17 - 19 , 2010 _ , pages 493502 , 2010 . benny godlin , tomer kotek , and johann a. makowsky .
evaluations of graph polynomials . in _ graph - theoretic concepts in computer science , 34th international workshop , wg 2008 , durham , uk , june 30 - july 2 , 2008 . revised papers _ , pages 183194 , 2008 .
johann a. makowsky .
connection matrices for msol - definable structural invariants . in _ logic and its applications , third indian conference , icla 2009 , chennai , india , january 7 - 11 , 2009 .
proceedings _ , pages 5164 , 2009 .
johann a. makowsky , udi rotics , ilya averbouch , and benny godlin . computing graph polynomials on graphs of bounded clique - width . in _
graph - theoretic concepts in computer science _ , pages 191204 .
springer , 2006 . | graph polynomials which are definable in monadic second order logic ( @xmath0 ) on the vocabulary of graphs are fixed - parameter tractable ( @xmath1 ) with respect to clique - width .
in contrast , graph polynomials which are definable in @xmath0 on the vocabulary of hypergraphs are fixed - parameter tractable with respect to tree - width , but not necessarily with respect to clique - width .
no algorithmic meta - theorem is known for the computation of graph polynomials definable in @xmath0 on the vocabulary of hypergraphs with respect to clique - width .
we define an infinite class of such graph polynomials extending the class of graph polynomials definable in @xmath0 on the vocabulary of graphs and prove that they are fixed - parameter polynomial time ( @xmath2 ) computable , i.e. that they can be computed in time @xmath3 , where @xmath4 is the number of vertices and @xmath5 is the clique - width . |
lattice gauge theory ( lgt ) can be formulated in many equivalent ways . the original wilson formulation is given in terms of group valued matrices on links as fundamental degrees of freedom @xcite .
the partition function reads @xmath3 \ } \ , \label{partfunc}\ ] ] where @xmath4 is the standard wilson action and the integral is calculated over the haar measure on the group at every link of the lattice .
several years later the dual representation for abelian lgt s was constructed in @xcite .
extensions to nonabelian groups have been proposed in ninetees in @xcite-@xcite .
a closely related approach is so - called plaquette representation invented in the continuum theory by m. halpern @xcite and extended to the lattice models in @xcite . in this representation
the plaquette matrices play the role of the dynamical degrees of freedom and satisfy certain constraints expressed through bianchi identities .
both dual and plaquette formulations are not so popular in the case of nonabelian models , probably due to their complexity . in the nonabelian case above mentioned constraints appear to be highly - nonlocal and this fact makes an analytical study of the model rather difficult . in spite of this complexity the plaquette formulation has certain advantages over the standard one @xcite : duality transformations , coulomb - gas representation , strong coupling expansion and mean - field approximation look more natural and simpler in the plaquette formulation .
nevertheless we believe the main advantage of this formulation lies in its applications to the low - temperature region .
let @xmath5 be a plaquette matrix in @xmath1 lgt .
the rigorous result of @xcite asserts that the probability @xmath6 that @xmath7 is bounded by @xmath8 uniformly in the volume .
thus , all configurations with @xmath9 are exponentially suppressed .
the gibbs measure at large @xmath10 is strongly concentrated around configurations on which @xmath11 .
this property justifies expansion of plaquette matrices around unity when @xmath10 is sufficiently large .
this is our first motivation to construct a low - temperature expansion of gauge models using the plaquette formulation .
the well - known problem of the standard perturbation theory ( pt ) can be shortly formulated as follows . when the volume is fixed , and in the complete axial gauge the link matrices perform small fluctuations around the unit matrix and the pt works very well producing the asymptotic expansion in inverse powers of @xmath10 .
however , in the thermodynamic limit ( tl ) system deviates arbitrarily far from the ordered state , no saddle point exists anymore and configurations of link matrices are distributed uniformly in the group space . that there are problems with the conventional pt was proven in @xcite , where it was shown the pt results depend on the boundary conditions ( bc ) used to reach the tl .
all this rises the question whether pt gives asymptotic expansion uniformly valid in the volume .
fortunately , even in the tl the plaquette matrices are close to unity , the inequality ( [ chestplaq ] ) holds and thus provides a basis for the construction of the low - temperature expansion .
our next motivation is to derive the dual representation of nonabelian models which could be used for their analytical study .
here we formulate such representation for @xmath2 lgt and compute dual form of the t hooft loop .
we also describe monopoles of the maximal axial gauge .
now we outline the derivation of our plaquette formulation ( all technical details will be given in a forthcoming publication @xcite ) .
consider a @xmath12 lattice with free bc on the link gauge matrices .
we make a change of variables in the partition function ( [ partfunc ] ) @xmath13 .
the partition function gets the form @xmath14 j(v_p ) \ , \ ] ] @xmath15 the last equations are rather formal , for we have to specify the order of multiplication of nonabelian matrices . to do that we define in a cube a vertex @xmath16 and separated by the space diagonal a vertex @xmath17 ,
see fig.[connector ] .
this assignment we extend for all neighbouring cubes and finally for the whole lattice .
the plaquette matrices @xmath5 we insert at the vertices @xmath16 or @xmath17 of each plaquette attached to @xmath16 or @xmath17 , correspondingly .
next , take a path connecting vertices @xmath16 and @xmath17 by three links , as shown in fig.[connector ] .
then the matrix @xmath18 entering the bianchi identity for the cube @xmath19 is of the form @xmath20 = \left ( \prod_{p \in a } v_p \right ) \
c \left ( \prod_{p \in b } v_p \right ) \
c^{\dag } \ , \
c = u_{1 } u_{2 } u_{3}^{\dag } \ .\ ] ] @xmath21 means an appropriately ordered product over three plaquettes of the cube attached to the vertex @xmath16 or @xmath17 .
the matrix @xmath22 defines a parallel transport of vertex @xmath17 into vertex @xmath16 and is called connector .
the connector @xmath22 plays a crucial role in the nonabelian bianchi identity .
its path can , in principle be chosen in an arbitrary way but the possible choices of @xmath23 , @xmath24 and @xmath22 must fit together .
we choose the structure of connectors as shown in fig.[connector ] . as is seen from fig.[connector ]
there are four types of elementary cubes with different types of connector pathes .
the next steps in the derivation are as follows : i ) .
map one of the space - like plaquette delta s to a delta - function for the cube @xmath25 with @xmath18 given in ( [ ordprod ] ) .
ii ) . choose the maximal axial gauge in the form @xmath26 to simplify connectors one can fix the dirichlet bc in the plane @xmath27 .
integrate out all link variables following @xcite . in what follows
we pass to the dual lattice where cubes become sites and so on .
the partition function takes the form @xmath28 \ \prod_{i=1}^4 \
\prod_{x(i ) } \ j\left ( v_x^{(i ) } \right ) \ , \ ] ] where @xmath29 runs over all sites with @xmath30-th type of connector pathes and @xmath31 is an @xmath1 delta - function given by @xmath32 where the sum over @xmath33 is a sum over all representations of @xmath1 , @xmath34 is character of @xmath33-th representation and @xmath35 . to write down exact expressions we choose coordinate system on the dual lattice as in fig.[connector ] .
then , e.g. the connector for the cube @xmath36 is @xmath37 @xmath38
in this section we construct the low - temperature expansion of gauge models in the plaquette formulation . in doing this
we follow the strategy developed for the two - dimensional principal chiral models in @xcite .
the weak - coupling expansion in both models bears many common features . in the abelian case they turn out to be practically identical , while the only , but essential difference in the nonabelian case is the appearence of connectors in gauge models .
the lattice pt in temporal gauge ( @xmath39 ) was constructed for abelian and nonabelian models in the standard wilson formulation by f. mller and w. rhl @xcite . in this gauge
the pt for nonabelian models displays serious infrared divergences in separate terms of the expansion starting from the order @xmath40 for wilson loops and from @xmath41 for the free energy .
although it is expected that all divergences are cancelled in gauge invariant quantities , we are not aware of any proof of the infrared finitness at higher orders of the expansion . here
we outline the low - temperature expansion for @xmath2 lgt ( all details and extension to other groups will be given in @xcite ) .
we want to expand the partition and correlation functions into asymptotic series whose coefficients are calculated over certain gaussian measure .
let us consider the standard parameterization for the @xmath2 link matrix @xmath42 $ ] , where @xmath43 are pauli matrices and introduce @xmath44^{1/2 } \
\ , \ w_x = \left [ \sum_k\omega^2_k(x ) \right]^{1/2 } , \label{wp}\ ] ] where @xmath45 is a site angle defined in ( [ ordprod ] ) and has the standard expansion @xmath46 in powers of link angles .
the first coefficient can be written down as @xmath47 @xmath48 for @xmath49 can be computed by making repeated use of the campbell - baker - hausdorff formula .
then , the partition function ( [ dualpart ] ) can be exactly rewritten to the following form @xcite @xmath50 \exp \left [ 2\beta\sum_l\cos w_l \right ] \prod_x \frac{w_x}{\sin w_x } \nonumber
\\ \prod_x \sum_{m(x)=-\infty}^{\infty}\int\prod_kd\alpha_k(x ) \exp \left [ -i\sum_k\alpha_k(x)\omega_k(x ) + 2\pi im(x)\alpha ( x ) \right ] \ , \label{pfwk}\end{aligned}\ ] ] where @xmath51 . in derivation of this representation
we have used the poisson formula . in order to perform the weak coupling expansion
we make the substitution @xmath52 , @xmath53 and then expand the integrand of ( [ pfwk ] ) in powers of fluctuations of the link fields .
introduce external sources @xmath54 coupled to the link field @xmath55 and @xmath56 coupled to the auxiliary field @xmath57 .
then one can get the following expansion @xmath58 m(h , s ) \ , \label{appa1}\ ] ] where @xmath59 are known operators .
up to exponentially small terms the generating functional @xmath60 is given by @xmath61 \ ,
\label{gffin}\ ] ] where we introduced link green functions @xmath62 and @xmath63 defined as @xmath64 @xmath65 these formulae allow us to calculate the weak coupling expansion of both the free energy and any fixed - distance observable .
extension of this expansion for @xmath1 lgt and for the wilson loop and other observables is straightforward .
let us give some comments on the property of the expansion and on the infrared problem .
one proves that the probability that @xmath66 links have large fluctuations like @xmath67 is suppressed as @xmath68 , @xmath22 is a constant .
it shows that large fluctuations are under good control since they are exponentially suppressed with @xmath10 , and this obviously remains true in the tl .
this property is very essential achievement of the expansion in the plaquette representation .
no such controll exists for the large fluctuations of the link gauge matrices in the tl .
it follows from the definition of the link green function ( [ gll1 ] ) that @xmath69 .
hence , with respect to the gaussian measure fluctuations of link variables are bounded like @xmath70 this property justifies the low - temperature expansion in powers of fluctuations of link ( plaquette ) variables , while there is no such justification for the expansion in terms of the original link variables the fluctuations of which are not bounded when @xmath71 . as is seen from the generating functional ( [ gffin ] ) it depends only on infrared finite green functions : link green functions are infrared finite by construction , while the green function for the free scalar field is finite in any dimension @xmath72 . in spite of all these nice properties , it is far from obvious that the expansion is uniformly valid in the volume .
the problem appears in the following way . in @xcite
we compute the first coefficient of the free energy expansion and prove that it coincides with the standard answer .
nevertheless , that computations show clearly that the source of the troubles are the connectors of the nonabelian bianchi identities .
the low - temperature expansion starts from the abelian bianchi identity for every cube and goes towards gradual restoration of the full identity with every order of the expansion .
thus , the generating functional contains an abelianized form of the identity without connectors .
high - order terms include an expansion of connectors which lead to appearence of infrared divergent terms .
consider , for example the term @xmath73 in the exponent of ( [ pfwk ] ) .
@xmath74 includes the following contributions from connectors @xmath75 , where sum over @xmath76 extends over all links belonging to the connector @xmath22 . for any finite @xmath77
one can always find such large @xmath10 that this sum is of the order @xmath78 . in the weak coupling expansion such terms
are treated as perturbations and should be expanded .
these terms are not infrared finite . and even though the final result is infrared finite it is not obvious that such terms may be considered as perturbation in the tl .
indeed , in the infinite volume limit practically all connectors become infinitely long , therefore it is not clear if the last sum is properly bounded in this limit and may be expanded perturbatively .
in this section we give the dual form of @xmath2 lgt which can be obtained from the plaquette formulation .
let @xmath79 be a site dual to the cube of the original lattice and @xmath76 be a link dual to the plaquette of the original lattice .
then it follows from ( [ dualpart ] ) , ( [ func2 ] ) and from the expression for the jacobian that in the case of the @xmath2 gauge group the partition function on the dual lattice can be written in the following form @xmath80 \prod_l \xi_l \ .
\label{pflsu2}\ ] ] the summation over @xmath81 corresponds to the summation over all irreducible representations of the @xmath2 group .
the sums over magnetic numbers @xmath82 correspond to the calculation of @xmath2 traces .
the index @xmath30 may run from @xmath83 to @xmath84 depending on the position of the original cube , @xmath77 is the linear extent of the lattice .
the link integral @xmath85 is given by @xmath86 where @xmath87 is a matrix element of @xmath33-th representation .
the integer number @xmath88 depends on the position of the plaquette on the lattice and shows how many times a dual link serves as connector in the bianchi identities .
consider now the t hooft line which in @xmath12 consists of a string @xmath89 of dual links between sites @xmath79 and @xmath90 on the dual lattice .
with the set @xmath91 of plaquettes dual to the string @xmath89 we define the action @xmath92 and the corresponding partition function @xmath93 with action ( [ taction ] ) . here
@xmath94 is center element of @xmath1 . by a change of variables @xmath95 for @xmath96
we get the following representation for the partition function @xmath93 @xmath97
\ \prod_{i=1}^3 \
\prod_{x(i ) } \ j\left ( v_x^{(i ) } \right ) \nonumber \\
\prod_{x(1)\ne x , y } \
j\left ( v_x^{(1 ) } \right ) \
j\left ( zv_x^{(1 ) } \right ) j\left ( zv_y^{(1 ) } \right ) \ , \label{sxypf}\end{aligned}\ ] ] since @xmath94 cancels from all bianchi constraints but those for the endpoints .
therefore , one can say that the bianchi constraint is violated at these endpoints .
it follows from these expressions that in the ensemble defined by the partition function ( [ pflsu2 ] ) the disorder operator becomes particulary simple @xmath98 \rangle \ .
\label{disopdual}\ ] ] comparing this expression with the partition function ( [ pfwk ] ) and remembering that @xmath81 is just the radial component of the auxiliary field @xmath57 we conclude that introducing sources like @xmath99 amounts to a simple shift in the corresponding sites for the integers @xmath100 in ( [ pfwk ] ) .
therefore , we may interpret the summation over @xmath100 as summation over monopole configurations which appear due to the periodicity of @xmath2 delta - function ( in close analogy with @xmath0 model ) . indeed , expressed in terms of elements of algebra the bianchi identity reads @xmath101 .
configurations @xmath102 we term monopoles of the maximal axial gauge .
investigation of these configurations and their contribution to the wilson loops will be presented in a separate publication .
wilson , k.g . , ( 1974 ) _
phys.rev . _ * d10 * p.2445 .
banks , t. , kogut , j. , myerson , r. , ( 1977 ) _ nucl.phys .
_ * b121 * ( 1977 ) p.493 .
anishetty , r. , cheluvaraja , s. , sharatchandra , h.s . ,
mathur , m. , ( 1993 ) _
phys.lett . _ * b314 * p.387 ; halliday , i. , suranyi , p. , ( 1995 ) _ phys.lett . _ * b350 * p.189 .
borisenko , o. , faber , m. , in bogolyubov , p. , jenkovszky , l. , ( eds . ) , new trends in high - energy physics , 2000 , kiev p.221 ; in lucha , w. , maung , k.m . , ( eds . ) , quark confinement and the hadron spectrum iv , 2002 , world scientific publishing , p.269 .
halpern , m.b . , ( 1979 ) _
phys.rev . _
* d19 * p.517 .
batrouni , g. , ( 1982 ) _ nucl.phys . _ * b208 * p.467 ; _ nucl.phys . _ * b208 * p.12 .
mack , g. , petkova , v. , ( 1980 ) _ annal.phys . _ * 125 * p.117 .
patrascioiu , a. , seiler , e. , ( 1995 ) _ phys.rev.lett . _ * 74 * p.1924 .
borisenko , o. , voloshin , s. , faber , m. , in preparation .
borisenko , o. , kushnir , v. , velytsky , a. , ( 2000 ) _
_ * d62 * p.025013 .
mller , v.f .
, rhl , w. , ( 1981 ) _ annal.phys . _
* 133 * p.240 . | we develop an analytical approach for non - abelian gauge models within the plaquette representation where the plaquette matrices play the role of the fundamental degrees of freedom .
we start from the original batrouni formulation and show how it can be modified in such a way that each non - abelian bianchi identity contains two connectors instead of four . using this representation
we construct the low - temperature expansion for @xmath0 and @xmath1 models on a finite lattice and discuss its uniformity in the volume .
next , we derive a dual representation for the t hooft loop in the @xmath2 model and describe monopoles in the maximal axial gauge . |
massless particles around the electroweak(ew ) scale determine physical laws observed around us .
the recent discovery of the higgs boson @xcite almost confirms the standard model(sm ) spectrum below about 1 tev .
obtaining such a massless spectrum theoretically has a long history under the name of the ` gauge hierarchy problem ' @xcite . since quarks and leptons have the chiral structure , they can be kept massless down to the tev scale .
however , the higgs boson is difficult to be kept massless in the sm framework .
the most studied solution of the gauge hierarchy problem has been supersymmetry(susy ) . even in the minimally supersymmetrized sm ( mssm )
, there is a problem known as the ` @xmath3 problem ' @xcite .
the mssm @xmath3 term , @xmath4 , is required to be at the electroweak scale , and the difficulty obtaining it at the tev scale is the problem .
recently , the permutation discrete symmetry @xmath5 has been used to obtain a massless pair of higgs doublets , @xmath6 and @xmath7 @xcite .
using a discrete symmetry to obtain the @xmath3 term is welcome in view of the gravity effects violating some ad hoc symmetries except the gauge symmetry @xcite .
indeed , it was argued in @xcite that the peccei - quinn(pq ) global symmetry @xcite broken at the intermediate scale has a severe fine tuning problem due to the gravity violation of the global symmetry .
if we consider a global symmetry , it must be an approximate symmetry .
discrete symmetries can not escape this gravity conundrum , but if it is a discrete subgroup of a gauge symmetry or dictated by string theory then it is safe from the gravity conundrum .
if the exact discrete symmetries are known , they can be helpful in obtaining an approximate global symmetry .
therefore , theoretical consideration of a global symmetry proceeds as follows : * firstly , at the energy scale much below @xmath8 , a global symmetry respected by the @xmath9 superpotential terms can be considered .
since it is an approximate symmetry , it must be violated by higher order terms beyond @xmath10 suppressed by the planck scale or the gut scale masses . *
second , the global symmetry in consideration from @xmath9 superpotential terms must be a part of a discrete symmetry respected by gravity so that the @xmath9 superpotential terms are not affected by the wormhole effects . *
third , for a specific global symmetry definition , all the superpotential terms up to @xmath11 may be needed for a unique definition .
the second and third steps are strong conditions .
this idea was presented in ref .
@xcite and here we realize the scheme explicitly . the regions for the discrete and global symmetries are shown in fig . 2 of @xcite . also , there is a common region which defines the approximate global symmetry from the discrete symmetry origin .
if we consider the sm gauge group below the gut scale , the rank of gauge symmetries we can count on are limited in the ultraviolet completed theory .
therefore , if there are too many singlets needed for the concrete definition of the global symmetry , the global symmetry may not be allowed from the ultraviolet completed theory . in ref .
@xcite , by obtaining a tev scale @xmath3 the approximate pq symmetry toward a strong cp solution was obtained .
the most curious cosmological observation is the extremely small dark endrgy(de ) of order @xmath12 @xcite .
the simplest form of de is the so - called cosmological constant(cc ) .
the cc appears as a real constant in the einstein equation @xcite and behaves as energy . on the other hand ,
if time - dependent homogeneous energy density evolves very slowly and stays almost constant during our observable time scale , it is called de .
so , de includes the cc . at present , it is fair to say that the cc problem is not understood @xcite . as in any attempts to derive a small de scale @xcite , here we assume that the cc is zero .
the anthropic bound @xcite is not a calculational scheme . .
at the point the thick arrow directs to , all equations of motion are satisfied and the cc is vanishing . ] the almost flat green - colored cartoon potential of fig .
[ fig : inflation ] is assumed to be that of a quintessential pseudoscalar field @xmath13 contributing to de .
it is different from the earlier terminology ` quintessential axion ' discussed in @xcite where the non - abelian anomaly breaks the global symmetry .
the pseudoscalar we consider in this paper is assumed not to couple to a non - abelian anomaly .
since the terminology ` axion ' is too much connected to the anomaly , in this paper we prefer to use the name ` @xmath14eudo@xmath15calar boson'(deps ) which does not couple to the qcd ( and hidden - sector if present ) non - abelian anomalies .
the corresponding approximate global symmetry is denoted as u(1)@xmath16 .
the easiest way to introduce the deps scale is just assuming a spontaneously broken u(1)@xmath16 potential whose height is about @xmath17@xmath18 above the theoretically favored cc value of 0 .
approaching to the current de scale . ]
but then we go back to an extreme fine tuning problem .
so , we propose to use suitable discrete symmetries toward obtaining our desired approximate global symmetry u(1)@xmath16 . even before the 1998 discovery of the accelerating universe ,
discrete symmetries were considered for obtaining some approximate global symmetry @xcite .
but with the 1998 discovery , the prospect of discrete symmetries as the basic tool for constructing global symmetries entered into the horizon of physics .
of course , the hypothetical discrete symmetry must satisfy the discrete gauge symmetry rule @xcite .
namely , we try to understand the de scale from two _ observed _ fundamental mass parameters , the planck mass @xmath8 and the tev scale @xmath19 , as envisioned in @xcite .
then , it has been argued that the suppression factor for the non - zero de potential with the intermediate scale of @xmath20 is of order @xmath21 @xcite .
since the gut scale can be considered as the planck scale multiplied by some o(@xmath22 ) parameters , we do not treat it as an independent scale .
the ew scale , @xmath3 and the tev scale are of the same order and we treat them as one scale
. the intermediate mass scale @xmath23 can be expressed as @xmath24 and we do not treat it as an independent scale . indeed , the scale @xmath3 can be related to the intermediate scale in some models @xcite . since the deps scale is independent of the qcd scale , we will not consider it here .
summarizing , let us introduce the parameters related to the deps in terms of two scales , @xmath25 with the possibility of some o(1 ) couplings multiplied .
if the deps potential of @xmath13 is expressed in terms of @xmath19 , @xmath13 must couple to the pairs of higgs doublets @xmath6 and @xmath7 .
if the combination @xmath26 carries a pq charge , @xmath13 must have a pq charge and we must worry about the anomaly u(1)@xmath16su(3)@xmath27 . to remove this anomaly , we introduce another u(1 ) symmetry u(1)@xmath28 so that one of the two linear combinations is free of the qcd anomaly .
namely , the deps scales ( deps mass and decay constant ) can not be treated independently from the qcd axion scale .
therefore , it is required to introduce two global u(1 ) symmetries to obtain the u(1)@xmath16 which is free of the u(1)@xmath16su(3)@xmath27 anomaly . as a consequence ,
two classes of singlet fields are introduced , @xmath29-type(@xmath30 and @xmath31 ) for the qcd axion and @xmath32-type(@xmath33 and @xmath34 ) for the deps .
if we introduce another confining force su(@xmath35 , an additional class of singlets will be needed not to introduce u(1)@xmath16su(n)@xmath36 . in susy models ,
the intermediate axion scale @xmath37 @xcite is considered as the confining scale of a hidden gauge group @xcite .
however , here we do not consider an extra nonabelian gauge group for simplicity and also because there may be some method to break susy without the extra nonabelian gauge group .
we attempt to obtain the deps of mass of order @xmath38 with the decay constant @xmath39 .
the qcd axion has been discussed with the @xmath5 symmetry @xcite .
another discrete symmetry is introduced to house one more approximate global symmetry for the deps , @xmath40 or @xmath41 @xcite . as a specific example , we consider the global symmetry @xmath42 .
the discussion is at the supersymmetric field theory level , even though it is strongly motivated by the orbifold compactification of the heterotic string . in sec .
[ sec : demo ] , we introduce the ` quintessence - flavor democracy ' to obtain the massless sm singlet fields , @xmath43 and @xmath44 . in sec . [
sec : depsv ] , we introduce the breaking terms of global u(1)@xmath16 symmetry and obtain the tiny height of the deps potential , @xmath18 . the u(1)@xmath16 breaking @xmath45-term dominates this breaking and renders the deps mass and the decay constant at @xmath46 and @xmath47 .
section [ sec : conclusion ] is a conclusion .
the first step toward understanding the de of the universe is to have a goldstone field @xmath48 where the superscript @xmath49 means the goldstone boson considering the lowest order interactions only .
the next step is to make it a pseudo - goldstone boson @xmath13 of fig .
[ fig : inflation ] so that higher order contributions are feeble enough for the tiny vacuum energy of fig .
[ fig : inflation ] .
so , we attempt to introduce an approximate global symmetry to have the pseudo - goldstone boson . since the global symmetry is approximate , this evades the gravity constraint that gravity does not respect global symmetries by the mother discrete symmetry @xcite .
the mother discrete symmetry is the one respected by gravity @xcite . in this scheme ,
the needed global symmetries u(1)@xmath28 and u(1)@xmath16 appear as accidental ones .
the goldstone field is massless at the tree level ( @xmath50 ) , whose vacuum energy scale is depicted at the bottom red line in the second column of fig .
[ fig : massscales ] . on the other hand ,
the expected energy scale in the fundamental lagrangian is @xmath51 which is depicted as the upper red band in the second column of fig .
[ fig : massscales ] .
the key point is how to obtain the bottom red line @xmath52 out of the gut or the planck scale energy @xmath51 .
a similar question on the @xmath3 problem of susy has been answered by introducing ` democracy ' for two pairs of higgs doublets @xcite . here
, susy is not a necessary ingredient , but is useful to simplify the couplings .
so , for the ` democracy ' idea , fermions in terms of the left - handed(l ) and the right - handed(r ) fields are introduced @xcite and we adopt susy for this purpose . for a natural introduction of the de scale , an extreme care is needed since any small correction at the gut scale @xmath1 , at the axion scale @xmath53 , or even at the electroweak scale @xmath19 , can easily outshoot the anticipated height of deps potential @xmath18 .
let us introduce two chiral singlets @xmath54 and @xmath55 with susy and each with two chiralities : l and r chiralities .
if these two singlets are not distinguished by any quantum number and geometry of the internal space , there must be the permutation symmetries , @xmath56 and @xmath57 .
the permutation symmetry @xmath58 is the permutation @xmath59 , each for the l and r fields .
the singlets , @xmath60 and @xmath61 are the singlet representations of @xmath56 , and the singlets @xmath62 and @xmath63 ( or represented by @xmath64 and @xmath65 ) are the singlet representations of @xmath57 .
then , the lowest order mass matrix @xmath66 of two singlets is @xmath67 where @xmath68 here , @xmath69 is the gut scale .
two eigenvalues of @xmath66 are @xmath69 and @xmath70 , whose scales are schematically shown as @xmath51 and @xmath52 , respectively , in the second column of fig .
[ fig : massscales ] .
the mass matrix @xmath66 is diagonalized to @xmath71 and the new bases are related to the old ones by @xmath72 or @xmath73 one may argue that starting with mass matrix ( [ eq : diagmassmatrix ] ) is just postulating two mass eigenvalues from the outset .
it is true but mass matrix ( [ eq : diagmassmatrix ] ) must have an underlying symmetry , which will become transparent by the inverse transformation , _ i.e. _ transforming eq .
( [ eq : diagmassmatrix ] ) to eq .
( [ eq : demomass2 ] ) . a similar line of reasoning has been adopted toward a solution of the @xmath3 problem @xcite . as stated in eq .
( [ eq : twoscales ] ) , we attempt to express the de scale in terms of the observed scales @xmath8 and @xmath19 . since @xmath19 is given by the vacuum expectation values(vevs ) of @xmath6 and @xmath7 , @xmath13 must couple to @xmath26 to be expressed in terms of @xmath19 . since @xmath74 carries a pq charge , @xmath13 couples generically to the qcd anomaly .
thus , we need another pseudoscalar , say the qcd axion @xmath75 , which also couples to @xmath26 such that one linear combination of them can be constructed , not carrying the qcd anomaly .
this anomaly free pseudoscalar is named again as deps @xmath13 . in this vein , the criteria of eq .
( [ eq : twoscales ] ) connect the qcd axion and deps with two global u(1 ) symmetries : u(1)@xmath28 and u(1)@xmath16 . since gravity does not respect global symmetries as commented in ref .
@xcite , we confine to discrete symmetries which are safe from gravity violation .
recently , simple criteria for obtaining @xmath76 and @xmath41 from string compactification have been given @xcite .
it relies on the vevs of the sm singlet fields . to use this simple method ,
the full spectrum of massless particles should be given in the compactification process .
in addition , due to two units of discrete charges of @xmath6 and @xmath7 , there can result a @xmath77 parity which guarantees a @xmath77-odd wimp particle . in this paper
, we introduce a discrete @xmath78 in addition to @xmath5 , to obtain the approximate u(1)@xmath16 and u(1)@xmath28 . here
, @xmath79 must be large enough so that it does not to allow the same u(1)@xmath28 and u(1)@xmath16 charges for two different type fields . here
, we choose @xmath80 so that the discrete symmetry is @xmath81 for the mssm fields plus heavy quarks and @xmath29-type and @xmath32-type singlets .
the permutation symmetry @xmath56 is the permutation @xmath59 for @xmath82 or @xmath83 or @xmath84 for @xmath85 , and @xmath57 is the permutation for @xmath86 or @xmath87 or @xmath88
. the relevant quantum numbers of the mssm fields , heavy quarks and @xmath29-type and @xmath32-type singlets are presented in table [ tab : quintcharges ] , including the electroweak(ew ) hypercharge @xmath89 . in the lower two rows , we present two global charges of u(1)@xmath28 and u(1)@xmath16 .
motivated by the string compactification , we are interested in matter which belongs to ( * 248 * , @xmath90 ) and ( @xmath90 , * 248 * ) of e@xmath91e@xmath92 . since *
248 * is not a singlet under e@xmath93 , it can not appear in @xmath94 as @xmath95 , _
i.e. _ any mssm signet matter field can not have a tadpole term in @xmath94 .
any component of * 248 * can be classified by eight u(1 ) charges of the cartan subalgebra of e@xmath93 .
any member , not belonging to the center of * 248 * , can not have all the vanishing u(1 ) charges .
suppose a non - center member @xmath32 of * 248*. certainly , not all the u(1 ) charges of @xmath96 for a non - center @xmath32 are vanishing , and @xmath96 is not an e@xmath93 singlet .
on the other hand , there is a possibility that @xmath97 can be an e@xmath93 singlet .
if @xmath32 is a center member , @xmath96 term can be present .
the above comment is checked for * 3 * of so(3 ) , @xmath98 where @xmath99 is a center member and @xmath100 are the non - center members .
so , @xmath96 in the superpotential is forbidden if @xmath32 does not belong to the center . for an su(@xmath101 )
subgroup of e@xmath93 , for example , the heavy quark @xmath102 and @xmath32 may belong to the same representation , and so may be the higgs field and @xmath29 , @xmath103 q^{(g)}\\[0.2em ] q^{(b)}\\[0.3em ] l^+\\[0.3em ] l^0\\[0.2em ] \xq\\ \cdot\\ \cdot \end{array}\right),~~ \left(\begin{array}{c } t^{(r)}\\[0.2em ] t^{(g)}\\[0.2em ] t^{(b)}\\[0.2em ] h_u^+\\[0.4em ] h_u^0\\[0.3em ] x\\ \cdot\\ \cdot \end{array}\right).\nonumber \end{split}\ ] ] the above grouping of fields is just an example .
but , here we introduce one different aspects for the @xmath29-type and @xmath32-type fields . we allow only one anomalous u(1 ) from the compactification . in this case , the anomalous gauge u(1 ) is broken near the string scale and a global symmetry survives down to the low energy scale .
let this be the u(1)@xmath28 because it carries the anomaly .
so , the u(1)@xmath28 breaking scale is at the intermediate scale @xcite . on the other hand ,
the u(1)@xmath16 is not anomalous and it must come from the non - anomalous u(1 ) gauge symmetry . in principle , the spontaneous symmetry breaking scale of this gauge u(1 ) is at the string scale .
then , the spontaneous symmetry breaking scale u(1)@xmath16 is determined by the vevs of u(1)@xmath16 charge carrying singlet fields . generically , these vevs can be raised to the planck scale , since we do not use @xmath104 but a phase field in matter fields as the deps .
now , the @xmath29-type and @xmath32-type fields survive down to the low energy scale where u(1)@xmath16 is already broken at @xmath105 and u(1)@xmath28 is not broken
. therefore , let us construct a model allowing the following features : * introduce @xmath106 .
introduce also a discrete symmetry @xmath107 which is large enough to introduce two approximate global symmetries , u(1)@xmath28 and u(1)@xmath16 . *
the u(1)@xmath28 has the color anomaly , and the very light axion is housed in @xmath108 and @xmath109 .
their vevs are at the intermediate scale .
* the u(1)@xmath16 does not have the color anomaly , and the deps is housed in @xmath110 and @xmath111 .
for this purpose , it is necessary to have heavy quarks @xmath112 and @xmath113 ( @xmath114 ) .
* there are three scales of vevs . out of two fundamental scales
@xmath1 and @xmath19 of eq .
( [ eq : twoscales ] ) , the intermediate scale @xmath115 and the deps parameters are derived . @xmath110 and @xmath111 must couple to @xmath116 and to @xmath101 pairs of heavy quarks .
we choose @xmath117 . *
the approximate global symmetries can be chosen into many directions . among these , u(1)@xmath28 carries the color anomaly .
for u(1)@xmath16 , the condition is the absence of u(1)@xmath16su(3)@xmath27 anomaly .
we take the u(1)@xmath28 breaking scale at the intermediate scale and the u(1)@xmath16 breaking scale at the planck scale .
* to have only one massless pairs through permutation symmetry , corresponding to the second column of fig .
[ fig : massscales ] , we introduce two pairs for each of @xmath118 , @xmath119 , @xmath120 , and @xmath121 .
therefore , the superpotential terms respecting the discrete symmetries of table [ tab : quintcharges ] are @xmath122 due to the @xmath5 symmetry , the singlet mass @xmath123 is of order @xmath1 and one massless pair , @xmath124 and @xmath125 , is obtained @xcite . without @xmath5 ,
the model can be more intricated to obtain the massless pair .
the masses @xmath126 and @xmath127 are of order the intermediate scale due to the couplings @xmath128 and @xmath129 .
@xmath130 is of order the kev scale due to the couplings @xmath131 .
the second line consists of dimension-3 @xmath94 , the third line consists of dimension-4 @xmath94 , the fourth line consists of dimension-6 @xmath94 , etc .
, all of which satisfy the discrete symmetry @xmath132 of table [ tab : quintcharges ] .
.the pq and quintessential charges from @xmath133 .
for the higgs fields , @xmath134 refers to @xmath6 and @xmath135 refers to @xmath7 .
all the chiral fields are left - handed .
the quantum numbers @xmath136 and @xmath137 are given for the light fields with superscript @xmath49 .
[ cols="^,^,^,^,^,^,^,^,^,^,^,^,^,^,^,^,^,^,^,^ " , ] if all the terms in eq .
( [ eq : wdiscrete ] ) are satisfied as the @xmath138 term does , then we are discussing a gauge symmetry .
the global charges can not be automatically satisfied in this way . to check the global symmetry we also introduced @xmath29 and @xmath139 in table [ tab : doublesu5 ] . under @xmath132 ,
the term @xmath140 can be present since it carries @xmath141 units of @xmath142 charge . for this term to be present from the u(1)@xmath143 gauge symmetry
, we have another conditions on the @xmath107 charges of @xmath29 and @xmath139 of table [ tab : quintcharges ] , @xmath144 because there are nonzero @xmath107 charges , the coefficients @xmath145 are completely determined by twelve charge conditions if e@xmath92 does not lead to a nonabelian gauge group .
99 # 1#2#3phys
. rep .
* # 1 * , # 2 ( # 3 ) # 1#2#3rev .
* # 1 * , # 2 ( # 3 ) # 1#2#3nucl .
* b#1 * , # 2 ( # 3 ) # 1#2#3phys .
* b#1 * , # 2 ( # 3 ) # 1#2#3phys .
* d#1 * , # 2 ( # 3 ) # 1#2#3phys .
lett . * # 1 * , # 2 ( # 3 ) # 1#2#3jhep * # 1 * , # 2 ( # 3 ) # 1#2#3jcap * # 1 * , # 2 ( # 3 ) # 1#2#3z .
* # 1 * , # 2 ( # 3 ) # 1#2#3euro .
j. * c#1 * , # 2 ( # 3 ) # 1#2#3j .
* g#1 * , # 2 ( # 3 ) # 1#2#3int .
j. mod .
phys . * # 1 * , # 2 ( # 3 ) # 1#2#3mod .
* a#1 * , # 2 ( # 3 ) # 1#2#3astrophys .
j. * # 1 * , # 2 ( # 3 ) # 1#2#3nature * # 1 * , # 2 ( # 3 ) # 1#2#3sov .
j. nucl .
# 1 * , # 2 ( # 3 ) # 1#2#3astrophys .
j. * # 1 * , # 2 ( # 3 ) # 1#2#3int .
j. mod .
phys . * # 1 * , # 2 ( # 3 ) # 1#2#3mod .
* a#1 * , # 2 ( # 3 ) # 1#2#3nature * # 1 * , # 2 ( # 3 ) # 1#2#3nucl . phys .
* b#1 * , # 2 ( # 3 ) # 1#2#3prog .
phys . * # 1 * , # 2 ( # 3 ) s. m. barr and d. seckel , ; m. kamionkowski and j. march - russell , [ hep - th/9202003 ] ; r. holman , s. d. h. hsu , t. w. kephart , e. w. kolb , r. watkins , and l. m. widrow , [ hep - ph/9203206 ] ; s. ghigna , m. lusignoli and m. roncadelli , ; b. a. dobrescu , [ hep - ph/9609221 ] . | democracy among the same type of particles is a useful paradigm in studying masses and interactions of particles with supersymmetry(susy ) or without susy .
this simple idea predicts the presence of massless particles .
we attempt to use one of these massless pseudoscalar particles as generating the cosmological dark energy(de ) potential . to achieve the extremely shallow potential of de , the pseudoscalar boson is required not to couple to the qcd anomaly .
so , we consider two pseudoscalars , one coupling to the qcd anomaly ( _ i.e .
_ qcd axion ) and the other not coupling to the qcd anomaly . to obtain these two pseudoscalars , we introduce two approximate global u(1 ) symmetries to realize them as the pseudo - goldstone bosons of the spontaneously broken u(1 ) s .
these global symmetries are dictated by a gravity respected discrete symmetry .
specifically , we consider an @xmath0 example , and attempt to obtain the de scale in terms of two observed fundamental mass scales , the grand unification scale @xmath1 and the electroweak scale @xmath2 . |
torres et al . ( 1997 ; hereafter t97 ) reported the discovery of a new bright ( @xmath0 ) quasar at relatively low redshift ( @xmath1 ) .
this object , called pds 456 , was discovered in the pico dos dias survey for young stellar objects , which uses optical magnitude and _ iras _ colours as selection criteria .
although somewhat fainter than 3c 273 , pds 456 lies close to the galactic centre and is seen through @xmath2mag of extinction ( t97 ) , and it is therefore intrinsically more luminous .
it is , however , radio - quiet , with @xmath3mjy ( griffith et al .
1994 ) and @xmath4mjy ( pds 456 can be identified with nvss j172819@xmath5141555 ; condon et al .
1998 ) , where @xmath6 is the flux density at @xmath7ghz .
in this paper we present optical and infrared observations of pds 456 .
we confirm the redshift found by t97 , and make a more accurate determination based on the unblended narrow emission line of [ fe ii ] @xmath81.6435@xmath9 m .
we also confirm the existence of significant galactic reddening . we compare the properties of pds 456 with those of 3c 273 , with which it has a number of similarities , and investigate the possibility that it is a radio - quiet analogue of the latter source .
we adopt @xmath10kms@xmath11mpc@xmath11 and @xmath12 . with this cosmology and our improved redshift determination ,
the proper distance to pds 456 is 1.01gpc .
_ jhkl@xmath13 m _ images of pds 456 were obtained using ircam3 on the united kingdom infrared telescope ( ukirt ) on ut 1997 august 24 .
standard near - infrared jittering techniques were used to allow flatfielding without the need to observe blank regions of sky .
the dark - subtracted images were scaled to the same median pixel value and median - filtered to produce a flatfield image which was divided into the individual frames .
these frames were then registered using the centroid of the quasar and averaged with bad pixel masking .
total integration times were 50s at _ jhk _ , 270s at @xmath14 , and 72s at @xmath15 .
conditions were photometric and flux calibration was performed using a solution derived from observations of a number of standard stars during the course of the night .
similar observations were also performed with the same instrument on ut 1997 september 16 .
near - infrared spectra of pds 456 in the @xmath16 and @xmath17 windows were taken using ukirt / cgs4 on ut 1997 august 12 and 10 , respectively .
conditions were photometric throughout and both spectra have a resolution @xmath18 and total integration times of 640s .
bad pixel masking and interleaving of the separate integrations was performed with the starlink cgs4dr software , and the remaining reduction was undertaken with the iraf data reduction package .
the spectra were taken through a 1.2-arcsec slit and extracted along a single 1.2-arcsec pixel .
the @xmath16 spectrum was corrected for atmospheric absorption and flux calibrated using the a2 star hd 161903 .
the @xmath17 spectrum used the f0 star sao 121857 as an atmospheric dividing standard and was flux calibrated using hd 225023 .
the spectra were wavelength calibrated using krypton ( @xmath16-band ) and argon ( @xmath17-band ) arc lamps and the r.m.s .
deviation from the adopted fit was @xmath19 .
an 813@xmath9 m spectrum of the source was taken with ukirt / cgs3 on ut 1997 august 30 as a service observation .
a 5.5-arcsec aperture was used with the low resolution grating , which gives a resolution @xmath20 .
triple sampling was employed and the total integration time was 3600s .
a spectrum was taken of @xmath21 aql ( altair ) at similar airmass , which was used for flux calibration using its ratio with the spectrum of @xmath21 lyr ( vega , assumed to be a 9400k blackbody ) from cohen & davies ( 1995 ) .
optical spectra were taken with the intermediate dispersion spectrograph ( ids ) on the 2.5-m isaac newton telescope ( int ) as service observations on the night of ut 1997 august 22 .
four grating settings were used to cover the entire optical spectrum from 33209150 at a dispersion of @xmath22pixel@xmath11 .
a 1.5-arcsec slit was used , and the seeing was reported to be 1.0arcsec .
the total integration time at each position was 500s , split into three separate exposures to facilitate cosmic ray removal .
wavelength calibration was achieved through the use of argon and neon lamps , and the r.m.s .
deviation from the adopted fit was @xmath23 for all wavelength regions except in the range @xmath24 , where the arc lines were saturated , and the deviations became as large as 1.5 .
we determined the flux scale using bohlin s ( 1996 ) spectrum of bd + 33@xmath252642 ( which was also observed with the same instrumental setup ) , available from the space telescope science institute .
the three exposures at each grating position were averaged together after masking out cosmic ray hits , and the spectra merged by using the regions of overlap to compute necessary greyshifts between the different grating positions .
shifts of up to 8% were needed to match up the flux scales , and it was noted that the fluxes in the different exposures taken at a single grating position differed by as much as 25% .
this strongly suggests the presence of clouds or significant slit losses , and as such the flux scale must be considered approximate .
the results of our aperture photometry are presented in table [ tab : photom ] . the flux calibration solutions for the night of 1997 sep 16 were slightly more uncertain than for the night of 1997 aug 24 , but we find no evidence for variability in any of the five filters at greater than @xmath26 significance .
this supports the lack of variability observed by t97 at optical wavelengths over a period of three weeks .
.infrared photometry of pds 456 in a 5-arcsec aperture .
[ cols="^,^,^,^ " , ] from this analysis , we measure the flux of the fe ii 37,38 multiplets as @xmath27wm@xmath28 and of the h@xmath29 line as @xmath30wm@xmath28 , three times larger than that determined by t97 .
we believe this discrepancy is due to an incorrect placement of the continuum level by t97 when performing their simple flux determination .
the value we obtain from our more detailed method is likely to be far more accurate .
the line fluxes are listed in table [ tab : fluxes ] .
note that for this table and future discussions , we have brightened our @xmath17-band spectrum by 25% to match our photometry , suspecting poor seeing and slit losses for the difference ; no such shift was needed for the @xmath16-band spectrum .
the equivalent width of h@xmath29 is fairly typical of nearby luminous quasars ( miller et al .
1992 ) and the ratio fe ii @xmath84570/h@xmath31 is not unusually strong .
we also have reason to suspect t97 s [ o iii ] @xmath85007 flux , which is blended with the @xmath85018 line of fe ii multiplet 42 , and which we fail to detect in our spectrum after removing the fe ii and h@xmath29 emission as described above .
t97 deblend these two lines and obtain gaussians of similar , narrow ( @xmath32kms@xmath11 ) , widths whose wavelengths are significantly different from those expected .
the mean wavelength of t97 s two deblended components is very close to the wavelength expected for the @xmath85018 line alone ( our revised redshift improves the agreement ) , and their combined flux is in excellent agreement with the flux we determine for this line alone .
the fe ii lines are also known to be broad ( we measure @xmath33kms@xmath11 ) , casting doubt on the narrow line produced by t97 s deblending .
we also fail to detect [ o iii ] @xmath84959 at its expected flux level .
we place an upper limit of @xmath34}~\lambda5007 ) < 2 \times 10^{-17}$]wm@xmath28 , corresponding to a rest - frame equivalent width of @xmath35 .
whilst the h i line ratios can be used to estimate the reddening , they are too uncertain to provide an accurate value , although they are broadly consistent with the @xmath36mag determined by t97 from the equivalent width of the na d 1 line , and the extinction maps of burstein & heiles ( 1982 ) .
we make an additional estimate of the extinction by comparing the optical spectrum of pds 456 to that of the quasar 3c 273 . as fig .
[ fig:3c273 ] shows , there is excellent agreement ( at least blueward of h@xmath21 ; we discuss the disagreement at longer wavelengths in the next section ) when pds 456 is dereddened by @xmath37 .
since 3c 273 is itself reddened by @xmath38mag ( burstein & heiles 1982 ) , we infer a total galactic extinction of 1.5mag towards pds 456 , and adopt this value in our later analysis .
in the previous section , we noted the similarity between the optical spectra of pds 456 and 3c 273 . in fig .
[ fig : optir ] , we present the dereddened optical infrared spectral energy distributions of the two quasars .
we use the data of rieke & lebofsky ( 1985 ) to correct the data at @xmath39 m , which is beyond the range of cardelli et al.s ( 1993 ) extinction law .
the similarity is once again striking , with the exception of the 10@xmath9 m flux , which is nearly a factor of two higher in pds 456 than 3c 273 , relative to the optical near - infrared data .
note that although these data are measured through different apertures , the quasar is more than 300 times brighter than an @xmath40 galaxy in the @xmath17-band ( gardner et al.1997 ) , and so the stellar contribution will be negligible .
there is clearly a problem with the continuum level redward of h@xmath21 , since it does not match the interpolation between the rest of the optical spectrum and the @xmath16-band spectrum .
this is not a simple flux calibration error affecting the reddest of the four int sub - spectra which can be corrected with a constant shift , since the match is excellent in the region of overlap with the adjoining subspectrum .
the dereddened optical continuum is well - described by a power law with spectral index @xmath41 .
this is bluer than the mean optical spectral index for quasars , but still within the observed range ( neugebauer et al .
the inflexion at @xmath42 m is a ubiquitous feature of quasar seds ( neugebauer et al .
1987 ; elvis et al .
however , the near - infrared bump usually extends to longer wavelengths , a power law often being able to fit the sed throughout the range 110@xmath9 m ( see neugebauer et al.1987 ) .
given the similarities between the optical and infrared properties of pds 456 and 3c 273 ,
it is natural to ask the question : is pds 456 a radio - quiet analogue of 3c 273 ? certainly the low equivalent widths of the optical forbidden lines and blue optical continuum suggest blazar - like properties .
a blazar nature should also be apparent at radio wavelengths , since the core emission should be strongly boosted by doppler beaming ( falcke , sherwood & patnaik 1996 ) , resulting in a flat radio spectrum which is more luminous than the general radio - quiet quasar population . to investigate this possibility ,
we have obtained vla a - array data at 1.4 and 4.85ghz .
a detailed discussion of the data will be presented in a future paper , but they confirm the nvss identification and flux , and reveal a steep radio spectrum . by extrapolating this spectrum to 8.4ghz ,
we predict a flux of 4.6mjy , which would cause pds 456 to lie on the optical radio luminosity relation for rqqs of kukula et al .
( 1998 ) .
the high optical luminosity and low forbidden - line equivalent widths of pds 456 are characteristic of the ` baldwin effect ' ( baldwin 1977 ) .
although several explanations for this effect have been advanced , including doppler - boosting ( browne & murphy 1987 ) , radiation from an accretion disc ( netzer 1987 ) , and reddening ( baker 1997 ) , these all incorporate an orientation dependence such that very luminous , very low equivalent width sources like pds 456 should be seen nearly pole - on .
the lack of a dominant , boosted radio core is therefore something of a mystery .
we have presented optical and infrared spectra of the nearby luminous quasar pds 456 .
we measure a redshift @xmath43 based on the forbidden line of [ fe ii ] @xmath81.6435@xmath9 m , but do not detect any other forbidden emission lines .
we detect at least three emission lines in the near - infrared which we are unable to identify .
the dereddened optical continuum is rather blue and very similar to that of 3c 273 . despite the similarities at optical wavelengths ,
observations reveal that pds 456 does not possess a strongly doppler - boosted radio core .
we defer detailed discussion of the radio observations to a later paper , wherein we will also present and discuss an x - ray spectrum of pds 456 .
the united kingdom infrared telescope is operated by the joint astronomy centre on behalf of the u. k. particle physics and astronomy research council , and was ably piloted by thor wold .
the isaac newton telescope is operated on the island of la palma by the isaac newton group in the spanish observatorio del rogue de los muchachos of the instituto de astrofisica de canarias .
the very large array is part of the national radio astronomy observatory , which is operated by associated universities , inc .
, under cooperative agreement with the national science foundation .
the authors thank the staff at ukirt and the int for taking the service observations presented in this paper , and duncan law - green for reducing the vla data . | we present infrared photometry and optical and infrared spectroscopy of the recently - discovered , extremely luminous nearby quasar pds 456 .
a number of broad emission features are seen in the near - infrared which we are unable to identify .
we measure a more accurate redshift from a narrow forbidden emission line and compare the optical infrared spectrum to that of 3c 273 .
the close similarity suggests that pds 456 is a radio - quiet analogue of 3c 273 , although radio observations do not support this idea . #
1to 0pt#1 galaxies : active quasars : individual ( pds 456 ) galaxies : distances and redshifts quasars : emission lines galaxies : photometry radio continuum : galaxies |
nearly half of the bolometric luminosity of the local universe is channeled through the mid- and far - infrared emission of galaxies ; this spectral region directly probes the youngest star - forming regions and their associated interstellar gas and dust .
however , existing infrared ( ir ) instruments have only begun to probe the full ranges in star formation rates , interstellar medium ( ism ) properties , and dynamical environments found in external galaxies .
the sirtf nearby galaxies survey ( sings ) is designed to characterize the infrared emission of present - day galaxies and probe the full range of star formation environments found locally , including regions that until now have been inaccessible at infrared wavelengths .
the observations include imaging and low - resolution spectroscopy of 75 nearby galaxies ( @xmath2 mpc ) , and high - resolution spectroscopy of their centers and a representative set of extranuclear infrared - emitting regions .
these data will be supplemented with an extensive library of ground- and space - based data at other wavelengths .
the primary scientific objective of sings is to use observations of these galaxies at infrared , visible , and ultraviolet wavelengths to obtain complete maps of the current star formation , that are free of most physical biases imposed by interstellar dust .
these will be used in turn to robustly model the extinction , heating , and infrared emission of the dust , and carry out a critical comparison and integration of ultraviolet , h@xmath1 , and infrared - based estimates of star formation rates ( sfrs ) in galaxies .
the infrared spectra will provide spectral diagnostics of the ionized , neutral , molecular , and dusty ism phases of individual star - forming regions , and will probe the interactions between young stars and their surrounding ism over a much wider range of physical and dynamical environments than can be probed in the galaxy .
in addition , sings will provide an archive of spatially - resolved images and spectra of nearby galaxies that will enable a rich set of follow - up investigations of star formation and the ism .
the galaxy sample and observing strategy are designed to maximize the long - term scientific return of the data for the sirtf user community as a whole .
the multi - wavelength archive , which will include sirtf observations as well as a set of @xmath0 , h@xmath1 , paschen-@xmath1 , ultraviolet , far - infrared and submillimeter , x - ray , co , hi , and radio continuum data , will provide a foundation of data and comparison samples for future sirtf general observer ( go ) observing projects by other investigators .
this paper is organized as follows . in
2 we describe the core science program for sings , and highlight some of the many archival applications of the data set . in
3 we describe the design and the properties of the galaxy sample , and 4 describes the sirtf observations themselves .
the ancillary data being obtained by our team and by other groups are described in 5 . as with all sirtf legacy projects the pipeline data products will be available immediately , with no proprietary period .
however , the sings team will also produce a set of enhanced value data products ( e.g. , mosaicked images , spectral data cubes ) , and software tools , and these are described in 6 . for a general description of sirtf and its instruments
we refer the reader to gallagher , irace , & werner ( 2002 ) .
information on the large - scale distribution of star formation in galaxies , and its coupling to the underlying properties of the ism is critical to a host of larger astrophysical problems the physical nature and origin of the hubble sequence , the structure and phase balance of the ism , the interpretation of observations of the high - redshift universe , the nature and triggering of starbursts and as a lynchpin for the physical understanding and modeling of galaxy formation and evolution .
nearby galaxies offer the unique opportunity to study this sfr
ism coupling on a spatially - resolved basis , over large dynamic ranges in gas density and pressure , metallicity , dust content , and other physically relevant parameters .
sirtf s unique contribution derives from its ability to probe the flow of energy from stars into the ism on scales ranging from the immediate environs of star - forming molecular clouds to the radiative transfer of diffuse radiation on kiloparsec scales ( helou 2000 ) . within star - forming regions , stellar radiation shining on the surrounding gas produces a series of ism components ,
all of which are probed in the infrared .
( figure 1 ) .
ionizing radiation produces an hii region that is traced by a suite of mid - infrared fine structure lines from species covering a wide range of ionization states . outside of the ionization front
the gas becomes optically thick to hydrogen - ionizing photons , but ultraviolet photons in the range 6 ev @xmath3 13.6 ev dissociate molecules , ionize atoms like c , si , and fe , and heat the gas via grain photoelectric heating to form a photodissociation region or pdr ( hollenbach & tielens 1999 ) .
deeper within the clouds , the warm molecular interface is directly observable via the pure rotational transitions of molecular hydrogen ( h@xmath4 ) .
dust grains are present throughout these phases and provide diagnostics of the physical conditions in the star - forming regions , through the thermal continuum emission and the aromatic emission bands .
much of this emission arises from dust grains in hii regions and pdrs , but dust in ambient clouds illuminated by the general interstellar radiation field ( isrf ) can also be important ( lonsdale & helou 1987 ) .
another energy channel mechanical energy transfer can be traced via shock - sensitive features such as h@xmath4 , [ siii ] , and [ feii ] .
in contrast to the study of the formation of individual stars , which has advanced dramatically over the past two decades , our theoretical understanding of star formation on galactic scales remains relatively immature .
the limitations begin at the very foundation , with our limited ability to measure and map the large - scale star formation rates ( sfrs ) in galaxies .
ground - based observations using h@xmath1 and other nebular recombination lines and direct observations of the ultraviolet continua of galaxies from space have provided high - resolution maps of star formation in hundreds of nearby galaxies ( kennicutt 1998a and references therein ) .
however the interpretation of these data is limited by interstellar dust , which absorbs roughly half of the emitted ultraviolet ( uv ) and visible radiation of typical star - forming galaxies , and more than 90% in extreme cases such as infrared - luminous starburst galaxies ( e.g. , calzetti et al .
1994 ; wang & heckman 1996 ; buat et al . 1999 , 2002 ;
sullivan et al .
2000 ; bell & kennicutt 2001 , calzetti 2001 ; goldader et al .
2002 ) . since the intensity of the star formation itself
is strongly correlated with the column densities of gas and dust ( kennicutt 1998b ) , the most extincted regions preferentially occur in the regions with highest sfr density . thus our picture of galactic star formation based on visible - uv observations alone is both incomplete and physically biased
. this bias may be even more important at high redshift , where intense and compact starbursts associated with the formation of galactic spheroids are prevalent ( e.g. , barger et al .
2000 ; giavalisco 2002 ) .
a similar physical dichotomy between uv - bright and ir - bright regions is seen on the scale of individual star - forming regions ( kennicutt 1998a , b ) .
the giant ob / hii associations in disks which dominate the visible star formation have mean emitting gas densities of order 10100 @xmath5 , neutral @xmath6 molecular column densities of order @xmath7 @xmath8 and visual extinctions of 03 magnitudes .
by contrast the regions that contain the bulk of star formation in ir - luminous starburst galaxies are located in circumnuclear environments , with typical densities of order @xmath9 @xmath5 , molecular column densities of @xmath10 @xmath8 , and visual extinctions of 51000 magnitudes ( kennicutt 1998a ) .
visible and uv diagnostics can not be applied reliably in these regions , and instead most of our information comes from ir and radio recombination and fine - structure lines . because of this observational segregation between the physical regimes that have been ( in)accessible in the uv visible and ir , the two sets of spectral diagnostics have not been fully tested against each other , and questions remain about whether we are interpreting the respective sets of spectra in a consistent manner .
the long - standing controversy over whether the initial mass function ( imf ) of young stars in ir - luminous starbursts is radically different from that observed in optically - selected star - forming regions is a prime example of this ambiguity ( e.g. rieke et al . 1993 ; lutz 2000 ; alonso - herrero et al .
an important step toward breaking this impasse is to apply a homogeneous set of ir spectral diagnostics across the entire range of star - forming regions , ranging from the most dust - obscured nuclear regions to objects in the disk that have traditionally been studied using optical spectra .
the infrared astrononical satellite ( iras ) survey detected the integrated dust emission from tens of thousands of nearby galaxies , and revealed a new class of infrared - luminous starburst galaxies ( soifer , neugebauer , & houck 1987 ) . since that time several groups have combined iras data for nearby galaxies with uv and/or h@xmath1 fluxes , to test the reliability of the far - infrared ( fir )
luminosity of a galaxy as a quantitative sfr tracer , and to quantitatively assess the effects of extinction on the uv and optical sfr tracers ( e.g. , helou 1986 ; lonsdale & helou 1987 ; rowan - robinson & crawford 1989 ; devereux & young 1990 ; sauvage & thuan 1992 ; buat & xu 1996 ; walterbos & greenawalt 1996 ; bell & kennicutt 2001 ; buat et al . 2002 ; dopita et al . 2002 ; kewley et al . 2002 ) .
the fir has proven to be a uniquely powerful tracer of the sfr in galaxies dominated by young stars and with high dust optical depths , where the infrared continuum provides a near - bolometric measurement of the stellar radiation output and sfr ( kennicutt 1998b ) . interpreting this emission in normal galaxies
has proven to be more difficult , however . there
the dust can also be heated by older stars , and the fractional contribution of this cirrus " component to the total fir emission is a strong function of galaxy properties and the sfr itself ( helou 1986 ; lonsdale & helou 1987 ; rowan - robinson & crawford 1989 ; walterbos & greenawalt 1996 ) . moreover
, the infrared emission may under - represent the output of regions with low dust content . despite these concerns recent work comparing the reddening - corrected h@xmath1-based and fir - derived
sfrs show a surprisingly good agreement over a wide range of galaxy types and sfrs , which suggests that at least for normal galaxies the effects of older stars and dust opacity on the fir are less serious than anticipated ( e.g. , dopita et al . 2002 ; kewley et al .
unfortunately it is difficult to model the effects of dust extinction and dust heating in a realistic way , because of the complex geometry of the dust and heating sources , and the absence of high - quality fir data .
the best spatial resolution achievable in the far - infrared iras images is of the order of an arcminute , and for all but the nearest galaxies this is insufficient to robustly model the dust heating and radiation reprocessing .
major inroads into addressing these problems have come from the the infrared space observatory ( iso ) , which provided higher - resolution imaging as well as spectroscopic observations of galaxies in the 3200 range ( kessler et al .
several surveys directed at nearby normal galaxies were carried out , including the iso key project on the interstellar medium of normal galaxies ( helou et al .
1996 ) , which obtained comprehensive imaging and spectroscopy for a sample of 60 galaxies with a wide range of infrared and optical properties ( dale et al . 2001
; malhotra et al . 2001 ; lu et al . 2003 ) .
other major programs included an imaging survey of 77 shapley - ames spiral and s0 galaxies ( bendo et al .
2002a , b , 2003 ) , surveys of the virgo cluster ( boselli et al . 1998 ; tuffs et al . 2002 ;
popescu & tuffs 2002 ; popescu et al .
2002 ) , and an imaging and spectroscopic survey of normal and barred spiral galaxies ( roussel et al .
2001a ) .
the superior spatial resolution of iso relative to iras provided valuable new information on the distribution of emitting dust and the associated star formation in galaxies , but this advantage was mainly limited to wavelengths shortward of 15 @xmath11 m , where the dust emission is dominated by band emission from small aromatic hydrocarbon dust grains ( e.g. , cesarsky et al . 1996 ;
helou et al .
2000 ; sturm et al .
2000 ; hunter et al . 2001 ; roussel et al . 2001a , b ) .
an example is shown in figure 2 , which compares an iso map of the nearby galaxy ngc 6946 ( a member of the sings sample ) with a groundbased h@xmath1 image ( dale et al .
it is clear that in the outer disk the h@xmath1emission and dust emission trace the same general star formation component .
however there are large local variations in the infrared / h@xmath1 intensity ratio , which are due to changes in dust opacity , radiation field intensity , and differences in the ages of the star - forming regions .
the emission from the central starburst is much stronger in the infrared ( relative to the visible ) , because of the much higher optical depth of dust in that region . because the local intensity of star formation itself tends to correlate with gas and dust column density ( e.g. , kennicutt 1998b ;
also see fig .
4 ) , the infrared emission provides a probe of the most intense starbursts that is virtually inaccessible in the visible and ultraviolet .
iso has also made a major contribution by producing mid- and far - infrared spectra for a sample of infrared - luminous galaxies , and for a few lower - extinction star - forming regions in disks ( e.g. sturm et al .
2000 ; thornley et al .
2000 ; helou et al .
2000 ; malhotra et al .
1997 , 2001 ) .
the richness of emission - line and emission - band diagnostics in this region of the spectrum is illustrated in figure 3 , which shows iso sws spectra of m82 and the circinus galaxy with the principal diagnostic lines labeled .
these include ionized tracers of stellar photoionized hii regions ( e.g. , [ neii ] 12.8@xmath11 m , [ neiii ] 15.6@xmath11 m , [ siii ] 18.7 , 33.5@xmath11 m , [ siv ] 10.5@xmath11 m ) , high - ionization tracers for application to agns , shocks , and very hot stars ( e.g. , [ oiv ] 25.9@xmath11 m , [ nev ] 14.3@xmath11 m ) , low - ionization tracers of shocks and pdr regions ( [ feii ] 26.0@xmath11 m , [ siii ] 34.8@xmath11 m ) , the s(0 ) and s(1 ) molecular rotational lines of h@xmath4 ( 17 and 28 @xmath11 m ) , and the pah dust bands in the 514 region .
spectroscopy of [ cii ] 158 in normal star - forming galaxies has provided insight into the microphysics of the ism .
galaxies with warmer far - infrared colors ( larger 60/100 @xmath11 m flux ratios ) show a smooth decline in the [ cii]/fir ratio ( malhotra et al .
2001 ) , which has been attributed to less efficient photo - ejection of electrons from increasingly charged small grains in more actively star forming regions .
this interpretation is supported by a strong correlation between [ cii ] line emission and 510 dust emission in star - forming galaxies ( helou et al .
2001 ) , because the latter is a measure of the small grain photoelectric heating rate . in terms of iso continuum spectroscopy
, we have learned that the mid - ir spectrum ( @xmath12 ) is nearly shape - invariant on the scale of galaxies , as a consequence of the single - photon heating nature of the pah dust ( draine & li 2001 ) .
the shape of the complete infrared spectrum can be parametrized by a single variable , the 60/100 @xmath11 m flux ratio ( helou et al.2000 ; dale & helou 2002 ; lu et al . 2003 ) . both of these discoveries bode well for interpreting high - z galaxies using ir spectral templates based on nearby galaxies .
sirtf promises a breakthrough in this field , because it will produce spatially - resolved infrared spectral energy distributions ( seds ) of galaxies with sufficient angular and spectral resolution to separate individual dust emission and heating components .
the sirtf images have sufficient sensitivity to map the thermal dust emission over most or all of the optical extent of spiral and irregular galaxies .
likewise the spectroscopic sensitivity of sirtf makes it possible to apply a consistent set of diagnostics across the local populations of galaxies and star - forming regions . to contribute optimally to these investigations
, sirtf must be used in a manner that is best adapted to its finite angular resolution .
we have therefore adopted a two - part strategy .
in the first part , 75 galaxies will be imaged completely with irac and mips , and partially mapped with low - resolution spectroscopy , to characterize the global infrared and star - forming properties of nearby galaxies , and explore their dependencies on type and other integrated properties . in the second part , we will use complete spectroscopic observations to physically characterize a set of discrete infrared - emitting sources .
these include the nuclear regions of each of the 75 galaxies , and observations of 75 extranuclear regions in @xmath1320 of the galaxies , divided between visible star - forming regions and infrared - bright objects discovered with our sirtf images , chosen to contain the complete range of physical characteristics .
the combination of these two data sets will fully characterize the infrared properties of normal galaxies , and enable a broad set of astrophysical applications .
a major focus of our analysis will be to test and develop improved diagnostics of sfrs in galaxies , based on a critical comparison of the ultraviolet , h@xmath1 , and infrared data ( e.g. , popescu et al.2000 ; roussel et al .
2001a ; bell et al .
the key to breaking the degeneracies that currently limit the modeling of extinction and dust heating is to map the re - radiated dust emission on angular scales that are comparable to those of individual infrared - emitting ( and star - forming ) regions . in most star - forming galaxies
the bulk of the dust radiation is emitted in the far - infrared , with a peak between 50 and 150 ( figure 1 and dale et al .
2000 ; dale & helou 2002 ; bendo et al .
hence a central element of our program is imaging of the sings galaxies at 24160 with mips , and sed mapping at 1099 with irs and mips . at the shorter wavelengths these maps will possess comparable or superior resolution to the isocam 15 maps illustrated in figure 2 , and even at the longest wavelengths we will obtain images with over 200 resolution elements for a median - size galaxy in our sample ( @xmath14 10 , where @xmath15 denotes the isophotal diameter from the rc3 : de vaucouleurs et al .
the irac camera on sirtf will allow us to measure the mid - infrared dust emission at @xmath132 resolution .
although this spectral region ( 58 ) is roughly a factor of ten shortward of the thermal dust emission peak , and the radiation is dominated by band emission from molecular aggregates or small grains , studies with iso reveal a surprisingly good correlation between the dust emissivity at these wavelengths and the sfr , over an extended range of physical conditions and spatial scales ( roussel et al .
we will test the reliability of the mid - infrared emission as a quantitative sfr tracer , as it offers the tantalizing prospect of mapping star formation in galaxies on arcsecond angular scales .
the products of this analysis will be maps of the total star formation , as well as maps of the uv visible extinction and separate distributions for the uv - bright and dust - obscured star formation .
these will allow us to delineate the broad trends in sfr as functions of local and global galaxy properties , and they will reveal any systematic differences between the distribution and amounts of obscured _ vs. _ unobscured star formation that may be present , as functions of radial location , metallicity , galaxy type , or dynamical environment ( e.g. , kennicutt 1983 , 1998a ; roussel et al .
since the infrared - bright regions preferentially trace the densest and youngest star - forming environments , such patterns must be present .
another area of investigation will be the combination of the sfr distributions with spatially - resolved co and hi maps to probe the form of the star formation law over the full range of ism environments found in nearby galaxies .
the potential of such an analysis is illustrated in figure 4 ( taken from kennicutt 1998b ) , which compares the disk - averaged sfrs and gas densities in normal galaxies ( symbol ) and infrared - selected starburst galaxies ( symbol ) .
there appears to be a single schmidt law ( index n @xmath13 1.4 1.5 ) which reproduces the sfrs across this range of regimes .
we will combine the spatially - resolved sfrs maps with gas density data from the berkeley illinois millimeter array survey of nearby galaxies ( bima song ) project ( helfer et al .
2003 ) and high - resolution hi maps ( see 5.5 ) to test whether this law holds on a point - by - point basis .
any such application is limited by the validity of the co emissivity as a molecular gas tracer , of course , and the dust masses derived from the sirtf sed maps should enable us to place some useful limits on the variability of the co / h@xmath4 conversion factor .
other applications will include direct mapping of the dust density and temperature distributions ( e.g. , dale et al .
2000 ; li & draine 2002 ) , construction of an improved radiative transfer model for dust in galaxies using the dirty code ( gordon et al . 2001 ; misselt et al .
2001 ) , and use of the unprecendented sensitivity of sirtf to search for dust in gas - poor galaxies , metal - poor objects , and in the halos of galaxies .
sirtf will also allow us for the first time to apply the powerful suite of mid - infrared spectral diagnostics to galactic ir - emitting components that span the full range of metallicities , densities , extinctions , and radiation field properties found in normal galaxies .
these observations provide the link between the observed infrared spectra of galaxies and the physical conditions in the emitting gas and dust , and are crucial inputs for interpreting the integrated seds and spectra of high - redshift objects and more luminous starburst galaxies .
they will also allow us to study the dependence of the physical and chemical properties of the dust ( e.g. , size distribution , composition ) on environmental conditions such as metal abundance and sfr . as a quantitative illustration of the sensitivity of sirtf for these applications
, we can apply the nominal pre - launch instrumental sensitivity estimates for a few applications of interest . for a galaxy at a fiducial distance of 3.5mpc ( the distance of the m81 group , a primary component of our sample ) , the spatial resolution of the telescope projects to linear dimensions ranging from 40pc ( 2 pixels , 2.58@xmath11 m ) to 100 , 300 , and 700pc ( @xmath16/d respectively at 24 , 70 , and 160@xmath11 m ) . at this distance a 10s set of mips integrations at 24 , 70 , and 160@xmath11 m will detect with 10-@xmath17 significance a star forming region with sfr = @xmath18 yr@xmath19 ( comparable to the orion core ) , with a typical cloud mass ( gas @xmath6 dust ) of order @xmath20m@xmath21 . likewise a 10s integration set with irac can detect the dust emission from @xmath1310@xmath22 m@xmath21 of ism illuminated by the average isrf .
these cloud masses correspond to limiting gas column densities over the respective resolution elements of order a few times @xmath23 to @xmath24 h @xmath8 , with deeper limits achievable with longer integration times . for high - resolution spectroscopy , irs would measure the [ neii ] 12.8 @xmath11memission line from a orion - luminosity region at 10-@xmath17 significance in a 30s integration , or the h@xmath4 emission from the s(0 ) and s(1 ) lines at 17 and 28 @xmath11 m of a @xmath25 m@xmath21 cloud at 100 k in 500s ( again at 10@xmath17 significance ) .
these numbers serve to illustrate the wide range of applications that will be possible either with the sings archive or follow - up go observations . when the irs spectra are combined with the ir seds and uv - visible spectra we will be able to carry out powerful cross - checks on the sfr and dust properties of the regions as diagnosed by the independent tracers .
specific investigations will include the use of the aromatic features in emission ( afe s ) to constrain dust heating and radiative transfer models ( e.g. , boselli et al .
1998 ; helou et al .
2000 , li & draine 2001 , 2002 ) , and as possible direct indicators of the sfr , as described earlier .
spectra of the ionized regions will be used to constrain systematic variations in the stellar temperatures and imfs in the galaxies , and to test the strong mid - infrared lines as quantitative sfr tracers ( e.g. , roche et al .
observations of the low - ionization lines ( e.g. , [ siii ] 34.8@xmath11 m , [ feii ] 26.0@xmath11 m ) and h@xmath4 rotational lines will be used to test for systematic differences in the physical conditions in the environments of circumnuclear and disk star - forming regions , and to test shock , hii region , and photodissociation diagostics based on these lines , over a much larger range of physical conditions than can be probed in the galaxy . as with all of the sirtf legacy science projects ,
the target samples and observing strategy were designed to support a wide range of scientific applications beyond the core program described in the previous section .
a comprehensive description of the types of archival applications of the sings data is beyond the scope of this paper , but a few examples will serve to illustrate the variety of potential applications .
one of our primary data products will be a complete galaxy sed library , including broadband integrated seds covering 0.4160@xmath11 m , and eventually with extended coverage over 0.15 850 for many objects , when our seds are integrated with observations from galex and jcmt ( see 5 ) .
a spectral library of the 150 discrete sources ( circumnuclear regions , extranuclear hii regions , extranuclear ir sources ) will also be produced , covering the 537 range , and with matching optical spectra or emission - line fluxes over the 0.360.7 range for most regions .
these data should be especially useful for interpreting data from other legacy programs ( e.g. , goods , swire ) , for interpreting submillimeter surveys of distant galaxies , and for planning future facilities such as alma , herschel , and safir .
the same data should be useful for modeling the composite emission from ultraluminous infrared galaxies ( ulirgs ) , and by helping to identifying the truly unique physical attributes of these objects ( not represented in the normal galaxies in our sample ) , and thus constrain their origins and evolutionary paths .
the irac images at 3.6 and 4.5@xmath11 m , when combined with groundbased near - infrared images , and correcting for hot dust emission at these wavelengths , can provide constraints on the stellar mass distributions in galaxies , and these are relevant to a wide range of problems including bar structures , spiral density waves , and better constraining the dark matter properties of galaxies , e.g. , by modeling the stellar mass to light ratio . the wide range of inclinations in the sample can also be used to study the extinction properties of disks as a function of wavelengths , and address the long - standing debate over the opacities of galactic disks at visible and near - infrared wavelengths ( e.g. , valentijn 1994 ; giovanelli et al .
1994 ; burstein 1994 ) .
the sings data will also provide several means of studying the spatial distribution of hot and cold dust in galaxies ( e.g. , lu et al .
the irac bandpasses have been selected to allow separation of aromatic emission band features from ( largely ) continuum regions , and thus it will be possible to construct high - resolution emission maps of the warm dust . at longer wavelengths
sirtf will offer unprecendented sensitivity to emission from the cold cirrus " component of dust , and these can be used to test for the presence of dust halos , streamers , or clouds around nearby galaxies ( e.g. , nelson et al .
1998 , alton et al .
1998 , trewhella et al .
2000 , stickel et al .
the mips images can also be used to place strong limits on the total ism ( gas + dust ) contents of early - type galaxies .
although it is not a main component of the sings core science program , the connection between agn activity and circumnuclear star formation activity can be explored in detail with these data .
the sample contains the full range of nuclear types , including hii region , liner , and seyfert nuclei .
the availability of high - resolution spectra in the 1037 region ( and low resolution at 514 ) will make it possible to compare the physical conditions and ionization properties of the regions from fine - structure line diagnostics with those derived from optical spectra .
the centers of the sings galaxies will be spectroscopically mapped at 15 positions ( 4 ) , providing some separation of nuclear and circumnuclear emission , within the limitations imposed by the spatial resolution of sirtf ( @xmath1329 fwhm at 537 ) . as a final example
, a wide range of gasdynamical processes , including large - scale shocks produced by bars , spiral density waves , and interactions are manifested in the mid - ir continuum and the shock sensitive [ siii ] , [ feii ] , and h@xmath4 features .
the irac imaging and low - resolution irs sed strips should isolate examples of these shocks , which can be followed up with pointed spectroscopy as part of the sirtf go program .
in order to maximize the archival value of the sings dataset , and to incorporate as diverse a population of galaxies and infrared emitting regions as possible , we adopted a physically - based approach to defining the samples .
these address the needs of the scientific program described in the previous section , and also provide representative samples for other astrophysical applications , as well as reference and control samples for future go observing programs . in order to achieve the scientific goals described above and at the same time to maximize the archival value of survey , we decided from the outset to observe as diverse a set of local _ normal _ galaxies as possible .
the size and precise nature of the sample , however , was heavily influenced by a set of technical considerations .
the principal limiting factor in the performance of sirtf for this type of science is its angular resolution , which ranges from about 2 with irac to 40 in the 160 imaging channel of mips . to maximize the linear resolution we need to study nearby objects .
we found that most of the range of normal galaxy properties could be covered by a sample with a maximum distance of 30 mpc ; the median distance of the sample is 9.5mpc ( for h@xmath26 = 70kms@xmath19mpc@xmath19 ) .
most of the irac and mips imaging channels cover a field of 5 square , so to make efficient use of the observing time we selected most of our galaxies to have sizes in the 515 range ( a few smaller galaxies were included in order to cover the full physical parameter space described below ) .
finally , the sample size was dictated by a balance between achieving a representative range of galaxy properties and practical limitations in observing time .
this led to a sample of 75 galaxies , requiring 512 hours of sirtf time . to construct a sample of candidate galaxies we adopted a methodology used in the iso key project ( helou et al .
1996 ) , and considered a 3-dimensional parameter space of physical properties : morphological type ( which also correlates with bulge / disk structure , gas fraction , and sfr per unit mass ) , luminosity ( which also correlates with galaxy mass , internal velocity , and mean metallicity ) , and fir / optical luminosity ratio ( which also correlates with the dust optical depth , dust temperature , and inclination ) .
we examined the distribution of nearby galaxies in this 3-dimensional space and selected objects that cover the full range of properties that ought to be detectable with sirtf .
specifically , we chose roughly a dozen galaxies in each rc3 ( de vaucouleurs et al .
1991 ) type ( e s0 , sa
i0 ) , which covered the full combination of luminosity and infrared / optical ratio .
figure 5 shows the distribution of sings galaxies in this parameter space .
the sample is very diverse , covering a factor of @xmath20 in infrared luminosity and @xmath27 in l@xmath28/l@xmath29 . when constructing the sample we also took pains to cover a representative range of other galaxy properties , including nuclear activity ( quiescent , starburst , liner , seyfert ) , inclination , surface brightness , co / hi ratio , bar structure , spiral arm structure ( grand - design vs flocculent ) , and environment ( i.e. , isolated galaxies , interacting galaxies , group members , and cluster members ) . our goal was not to cover a representative range of all of these secondary properties ( impossible with a sample of only 75 galaxies ) , but rather to avoid introducing an unintended bias into the sample with respect to any of these properties .
the final selection of galaxies also took into account practical factors such as galactic and ecliptic latitude .
we excluded most galaxies near the galactic plane to avoid high background at long wavelengths ( from interstellar cirrus emission ) , and to avoid a high density of foreground stars at the shorter wavelengths .
we also avoided concentrating too much of the program near the ecliptic equator , where the zodiacal background is high and scheduling of observations is more constrained . however
some important regions lie near the ecliptic equator ( e.g. , virgo cluster , leo groups ) , so some observations in this region were unavoidable .
table 1 summarizes the range of properties in the sings sample , and table 2 gives a complete listing of the galaxies and some of their relevant properties .
although our sample comprises representative cross section of galaxies found in the local supercluster , it does not include the absolute extremes in properties that can be found over larger local volumes . for example , we did not include any examples of ultraluminous infrared galaxies ( uligs ) , luminous agns , cd galaxies , dwarf spheroidal galaxies , the most extreme metal - poor dwarf galaxies , or luminous low surface brightness galaxies .
this was either because the objects would not be spatially resolved at the longest wavelengths measured by sirtf ( e.g. , uligs ) , or because ample samples of such galaxies already are included as reserved observations in the guaranteed time observers ( gto ) program . in order to maximize the science return on the sirtf observing time we chose to avoid duplication , in favor of eventually integrating any needed gto data into future analyses .
the deliberate emphasis on constructing a physically diverse sample also means that sings should not be regarded as a representative , randomly drawn sample of the local universe , in either a volume - limited or flux - limited sense .
this point in illustrated in figure 6 , which compares the distribution of 60 luminosities of the sings galaxies with the local infrared luminosity function from strauss & huchra ( 1988 ) .
we deliberately constructed the sings sample to sample roughly equal logarithmic bins in luminosity , and this is reflected in the flat distribution in figure 6 . as a result
the number of ir - faint galaxies ( mostly dwarfs ) is underrepresented relative to their actual space densities .
likewise the sample lacks any ultraluminous objects , as mentioned earlier .
this selection should be borne in mind when interpreting results from the sings sample as a whole , and particularly when using the survey as a control sample for follow - up go studies .
nevertheless the sample covers the range of galaxy types and star formation environments that are responsible for @xmath3095% of the total infrared emission in the local universe ( sanders & mirabel 1996 ) .
although it was impractical to construct a fully volume - representative sample for sings , we decided to incorporate one complete luminosity - limited subsample into the survey , namely the m81 group ( tully 1987 ; karachentsev et al .
our sample includes including m81 , m82 , ngc 2403 , ngc 2976 , ngc 4236 , ic 2574 , ho i , ho ii , ho ix , ddo 53 , ddo 165 , m81 dwa , and m81 dwb .
four other group members , ngc 2366 , ngc 3077 , ugc 4486 , and vii zw 403 are already in the reserved observation catalog , and those data eventually will be available to complete coverage of the group .
the m81 group offers the advantages of proximity ( d @xmath31 3.5 mpc ) , high visibility for sirtf , and a diverse galaxy population , ranging from extreme dwarfs ( @xmath32 for m81 dwb ) to a large spiral ( m81 ) and the nearest example of an infrared - luminous starburst galaxy ( m82 ) .
the spectroscopic sample is comprised of the nuclear regions of 74 of the 75 sings galaxies ( the center of m82 is a reserved target , to be mapped by the irs instrument team ) , and 75 extra - nuclear regions .
the extranuclear targets are divided between @xmath1340 optically - selected ob / hii regions and @xmath1335 infrared - selected regions , most of which will be chosen from the sings images themselves . a similar physically - based approach has been used to define the extra - nuclear target sample .
the optically - selected hii regions are located in nine of the nearest sings galaxies , and were chosen to span wide ranges in metal abundance ( 0.1 3 @xmath33 ) , extinction - corrected ionizing luminosity ( q@xmath34 photons
s@xmath19 ) , visual extinction ( 04 mag ) , radiation field intensity ( 100-fold range ) , ionizing stellar temperature ( @xmath35 = 35 55 kk ) , and local h@xmath4/hi ratio ( @xmath360.1 to @xmath3010 ) , as inferred from co and 21 cm maps .
we have restricted the optically - selected hii regions to those with extinction - corrected @xmath37 w m@xmath38 ( corresponding to @xmath39 w m@xmath38 ) , to ensure adequate detection of the main fine - structure lines in a reasonable exposure time .
table 3 lists the current candidate spectroscopic targets along with their relevant physical properties .
optically - selected hii regions cover only part of the full population of bright star - forming regions and infrared sources , so it is important to independently select a subsample of extranuclear targets from infrared maps .
we have used published isocam mid - infrared maps to select some of the target regions listed in table 3 , and will select at least 30 additional targets from the sings images , on the basis of infrared luminosity , mid - infrared colors , and infrared - optical flux ratios .
this will ensure a fully representative sampling of the sings extranuclear source population as a whole , including totally optically - obscurred objects that would never have been selected at other wavelengths .
in addition a number of infrared - bright circumnuclear regions will be included in the sings observations of the centers of these galaxies .
in keeping with the archival design of the sings observing program , we chose to maximize the homogeneity of the imaging and spectroscopic dataset across the 75 galaxy sample .
our basic observing strategy is depicted in figure 7 , which shows an isocam image of a sings target , ngc 6946 , on three scales with the footprints of the mips , irac , and irs observations overlaid . for each galaxy full - coverage imaging
( to @xmath40 ) will be obtained at all 7 irac and mips channels ( 3.6 , 4.5 , 5.8 , 8.5 , 24 , 70 , 160 ) , with sufficient additional sky coverage to allow for a robust background subtraction .
the coarse spectral sampling of the broadband mips and irac maps by themselves is insufficient to constrain the infrared sed shapes of galaxies , much less diagnose the temperature distributions and properties of the emitting grains , as illustrated in figure 8 .
therefore we are supplementing our imaging with irs 1440@xmath11mspectral maps and mips 5299@xmath11 m sed mode strip scans in all 75 galaxies .
these will permit robust modeling of the dust continuum and heating over a several square arcminute slice of each galaxy , and with the images help constrain the spectral shapes throughout the disks .
additional spectral maps will be obtained at 514 and 1037(at high resolution ) for the centers of every galaxy except m82 , and for the 75 extra - nuclear regions described earlier .
the remainder of this section describes the sings observing strategy in detail .
this information is provided in part for the benefit of prospective sirtf proposers , who may find it useful for planning their own imaging or spectroscopic observations of nearby galaxies or other bright extended sources .
a summary of the observations is given in table 4 ; readers who are only interested in the data products may wish to skip the detailed discussion .
a majority of our galaxies are larger than the 5 format of the irac detectors , so we will construct mosaics from multiple pointings to provide the necessary field coverage . a generous overlap ( 50% ) between adjacent exposures will provide robust cross - registration of the images , determination of bias drifts , and cosmic ray rejection .
after one series of exposures we will have observed all locations in the central part of the mosaic four times . because the four irac detectors map two non - overlapping fields of view , mosaics of the larger galaxies will include regions above and below which are only observed in two of the four channels .
these regions are located well outside of the galaxies , and will be used to measure the sky background level .
the sensitivity of irac is sufficiently high to saturate the detectors in the inner regions and in bright spiral arms , where 7 surface brightnesses of up to hundreds of mjy sr@xmath19 are expected ( dale et al .
1999 ) . to address this problem
we will observe the sings galaxies in a high dynamic range mode , in which an additional short exposure ( 12s ) is obtained .
this will allow us to correct for any saturated pixels in the longer exposures .
we will image in all four irac bandpasses ( 3.6 , 4.5 , 5.8 , 8.0 ) to at least the full optical radii ( @xmath41 ) of the galaxies .
the four 256@xmath42256 element detectors observe adjacent regions of the sky simultaneously , each covers a region of 5@xmath425with 12 pixels .
galaxies with @xmath43 will be observed in a pair of four - position dithered exposure sets .
dithered exposures will enable us to recover somewhat more spatial information in the reconstructed images , as described in 6.1 . in order to facilitate the detection and removal of asteroids ( a significant contamimation for targets near the ecliptic ) , our observations
will be carried out in two sets of exposures separated by 124h .
a majority of our galaxies are larger than the 5 format of the irac detectors , so we will construct mosaics from multiple pointings to provide the necessary field coverage . a generous overlap ( 50% ) between adjacent exposures will provide robust cross - registration of the images , determination of bias drifts , and cosmic ray rejection .
after one series of exposures we will have observed all locations in the central part of the mosaic four times . because the four irac detectors map two non - overlapping fields of view , mosaics of the galaxies will include regions above and below which are only observed in two of the four channels .
these regions are located well outside of the galaxies , and will be used to measure the sky background level . like the smaller galaxies , each mosaiced galaxy will be observed twice to allow for the detection of asteroids .
so there will be a total of eight exposures for all locations in the central part of the mosaic . in keeping with the overall design strategy of the sings project we will obtain uniform exposure times ( per detector field ) across the galaxy sample ,
apart from small variations within individual mosaics depending on how many times a given region of sky is covered by a particular detector .
the baseline integration time per pointing ( 30s ) was chosen as a compromise between providing acceptable signal / noise across the galaxies and keeping the total observing time to a practical level ( 45h ) .
for moderate zodiacal backgrounds ( the dominant source at these wavelengths ) we expect a limiting 1-@xmath17 sensitivity of 0.04 mjy sr@xmath19 , which is comparable to the expected average 7 surface brightness of a typical sings galaxy at r@xmath44 ( dale et al .
the signal - to - noise ratio improves at smaller radii and shorter wavelengths .
for example , we expect to get a signal - to - noise ratio per pixel of @xmath135 for detecting stellar photospheres at r@xmath44 .
since the zodiacal background varies by factors of several over the sky , the limiting sensitivities are expected to vary by about @xmath454050% across the sample
. however these variations will be small in comparison to the intrinsic variations in infrared emission among the galaxies themselves .
the sensitivity of irac is sufficiently high to saturate the detectors in the inner regions and in bright spiral arms , where 7 surface brightnesses of up to hundreds of mjy sr@xmath19 are expected ( dale et al .
1999 ) . to address this problem
we will observe the sings galaxies in a high dynamic range mode , in which an additional short exposure ( 12s ) is obtained .
this will allow us to correct for any saturated pixels in the longer exposures .
each galaxy in the sample will be imaged in all 3 mips bands ( 24 , 70 , and 160 ) .
we set as a goal a minimum s / n = 3 at r@xmath44 in all three bandpasses .
to estimate the surface brightness levels required , we used the radial profiles of 11 nearby , face - on galaxies observed with the iras satellite , and processed with the hires algorithm to provide maximum spatial resolution ( auman , fowler , & melnick 1990 ; kerton & martin 2001 ) .
these profiles were then used to determine the 25 and 60surface brightness at @xmath41 in each galaxy .
this exercise yielded average values of 0.21 mjy sr@xmath19 at 25 and 0.51 mjy sr@xmath19 at 60 , in very good agreement with the mid - infrared estimates above , when extrapolated to longer wavelengths using typical galaxy seds ( dale et al .
we adopted these values as the 3@xmath17 sensitivity goals at 24 and 70 .
typical galaxy surface brightnesses at 160 are comparable to or greater than those at 60 ( engargiola 1991 ; bendo et al .
2002a ) , so we also set a sensitivity goal of 0.51 mjy sr@xmath19 at 160 .
the 3@xmath17 goals above are equivalent to @xmath46 point - source sensitivities of @xmath47jy , 1.9 mjy , and 6.1 mjy , respectively . as with the irac observations , variations in zodiacal background ( at 24 ) and interstellar background ( at 70 and 160 ) will affect the actual results , but even in the high background regions we expect to achieve these sensitivity targets within a factor of two .
observations of galaxies at 70 and 160 impose special challenges for observations and data processing with mips , because of the extended low surface brightness structure in the sources , the difficulty of separating physical structures from transient features in the ge : ga detectors , and foreground structures such as interstellar cirrus in the milky way galaxy .
for that reason we describe the design of our observing and data analysis strategies in some detail , in hopes that it may aid future sirtf users in designing their observing programs . at the wavelengths covered by mips our observations
will be confusion limited , and the sensitivity of the detectors is sufficient to detect most regions in exposure times of seconds .
consequently we have chosen to use short integration times ( 34 s ) , and multiple integrations on each region , to minimize saturation effects and ensure a high level of redundancy in the data . for expected average background levels ( young et al .
1998 ) , the nominal sensitivities given above are achieved in 1 photometry cycle at 70 and 160 , and is within 10% of being achieved after 4 cycles at 24 .
multiple photometry cycles ( 23 , depending on wavelength ) will be obtained to ensure sufficient redundancy for removal of detector transients , meet the sensitivity goals , enable super - resolution capability ( 160 ) , and produce consistently high - quality images . as with the irac imaging the observations
will be divided into two sets offset by 1.5 160 pixels , and separated by at least 3 h to detect asteroids .
the mips instrument enables imaging to be obtained either in a conventional photometry " mode , or in a
scan map " mode , in which a scanning mirror allows for continuous integration over large areas of sky .
we will use the small - field photometry modes for sources smaller than about d@xmath48 , and large - field photometry modes for sources up to d@xmath49 .
larger galaxies will be imaged using the mips scan map mode . to improve sampling and mitigate the effects of bad pixels at 160 each scan leg
will be offset from the previous one by one - half the array width , using the @xmath50 cross - scan step in both the forward and reverse directions .
this imaging strategy ensures that each point on the sky will be observed by at least 40 ( 4 at 160 ) independent pixels . due to the ge
: ga transient behavior , we feel that 4 observations are the minimum required to generate high - quality data .
we will perform a small raster of sed mode observations ( 5299 ) in a radial strip in each galaxy , designed to overlap with the irs low - resolution strip scans ( below ) . in most galaxies ( d@xmath51 ) , this raster will consist of 7 positions , overlapped by half the slit width to cover in full a @xmath52 region . in galaxies larger than 10 , a second , similar set of sed observations will be added to produce a @xmath131@xmath53 map .
scheduling constraints are used to ensure that these strips be oriented as closely as possible to the irs spectral maps , so we will cover the same regions in the galaxies at 1440 and 5299 .
we used the radial profiles of nearby galaxies discussed above to estimate the typical surface brightnesses at r = 24 , and computed the integration times required .
the fainter regions correspond to @xmath131.2 mjy sr@xmath19 , which implies a @xmath46 sensitivity of 4.5 mjy on a point source to achieve @xmath54 sensitivity at that surface brightness .
this is achieved in 4 cycles of 10 s integrations , so we will perform two cycles at each map position so that the half - slit - width overlaps provide the total integration time required .
the irs observations consist of : ( 1 ) low - resolution radial strip maps of all 75 galaxies in the long - low channel ( 14.240.0@xmath11 m ) ; ( 2 ) targeted spectroscopy of the centers of all 75 galaxies and 75 extranuclear sources in a subset of galaxies , at 5.314.2@xmath11m(low resolution ) , and 10.037.0@xmath11 m ( high resolution ) .
we will obtain 0.9-wide spectral maps for all galaxies from 14.2 to 40.0@xmath11 m using the irs low resolution mode , maximally overlapping the mips sed strip .
these irs strips will extend radially to about 0.55@xmath41 , where the average surface brightness is @xmath551 mjy sr@xmath19 ( dale et al .
2000 ) . to minimize the effects of cosmic rays and bad pixels , we will spatially quadruple - sample by executing half - slit ( length and width ) steps parallel and perpendicular to the long - low slits .
since the irs short - low subslits are about one arcmin long and are almost perpendicular to the long - low slits , we will complement the 1-wide long - low strip maps with short - low maps ( [ sec : targeted_low_res ] ) , to obtain complete wavelength coverage from 5.3 to 40.0 @xmath11 m in the nuclear regions . again using the @xmath56+_iso_-based spectral model of dale et al .
( 2001 ) , we find that exposure times of 30 seconds give s / n @xmath57 for a 15 @xmath11 m surface brightness of @xmath131 mjy sr@xmath19 , for the cirrus - like seds in the outer disks .
low - resolution spectral maps from 5.314.2@xmath11 m , nyquist - sampled over a 03@xmath4209 region , will be obtained for the centers of each galaxy and the 75 extranuclear regions .
this region of the spectrum is especially valuable for measuring the aromatic dust emission bands from small grains ( including pah species ) .
since the galaxy centers will also be covered by the 14.240.0@xmath11 m and 5299@xmath11 m sed strips , we will have nearly complete spectral coverage of these regions .
moreover , the 55 sub - slit length will enable useful observations of extended sources , and the serendipitous regions simultaneously covered by the other sub - slit will provide data on the local background and quiescent disk regions .
we will use exposures of 14s for the nuclei and 60s for extranuclear regions .
based on the range of observed mid - infrared surface brightnesses for the _ iso _ key project galaxies ( @xmath58 mjy sr@xmath19 ) , we expect s / n @xmath59 5 in the core of each target . as for the low - resolution strip maps , we will map targeted regions with high spectral resolution , using a 5@xmath423 grid pattern of half - slit length and half - slit width steps , as shown in the righthand panel in figure 7 .
most of our targets are spatially extended , with significant emission structure over scales comparable to the sizes of the irs short - high aperture . as a result
it would be very difficult to obtain accurate spectrophotometry across the entire wavelength range ( 1037 ) with a single pointing observation , because the size of the point - spread - function ( psf ) varies by so much over this wide spectral range .
our spectral mapping strategy will provide : 1 ) adequate spatial coverage to obtain accurate line ratios over the same physical region in the sources ; 2 ) s / n @xmath57 in the principal diagnostic lines ( below ) ; 3 ) flexible scheduling and roll angle .
high - resolution observations with the irs spectrograph are covered in two echelle modes , a short - high " mode at 1019.5 and a long - high " mode at 19.337 .
our short - high spectral maps will cover a total area of 157@xmath42236 , while the long - high maps cover an area of 335@xmath42448 .
the larger area in the long - high observation is necessitated by the larger psf at the longer wavelengths .
our basic unit of coverage for the nuclei and hii regions is well - matched to the area covered by the short - high slit .
we expect to achieve 1 @xmath17 noise levels for unresolved emission lines of approximately @xmath60 w m@xmath38 in the short - high and long - high data , except for the region beyond about 32 @xmath11 m which may be a factor of 24 noisier .
thus , emission lines at the @xmath61 w m@xmath38 level , similar to what we expect for [ neii ] and [ siii ] , should be detected at s / n @xmath59 510 over the entire mapped area .
higher s / n is expected in the centers of the mapped regions , where we benefit from aperture oversampling .
the total sings observing program is 512 hr , divided roughly equally between irac and mips imaging , irs and mips low - resolution spectral mapping , and the targeted irs spectroscopy ( table 4 ) .
it is important to bear in mind that all of the legacy observing plans will be re - evaluated after the launch and commissioning phase of sirtf , and are subject to revision should there be significant variances in the flight performance of one or more of the instruments .
the sings observing strategy was deliberately designed to be robust against modest variances in instrumental performance , so unless the instrumental sensitivities or capabilities changes substantially , the program will still be carried out as described above . for a subset of ir - faint sources , including most of the e
s0 galaxies and some of the dwarf irregular galaxies , we have deferred scheduling the spectroscopic observations until after the irac and mips imaging have been obtained , so we can confirm the detection of a minimum threshold infrared surface brightness to make spectroscopy feasible .
detailed information can be found on the sings website ( url : http://sings.stsci.edu ) .
the scientific value of the sirtf data will be enhanced by a large set of ancillary multi - wavelength observations that are being obtained and assembled for the sings galaxies .
many of these data are being obtained by the sings team , and will be archived together with the sirtf data products ( 6 ) .
our project also has stimulated other groups to undertake complementary " surveys of the sings sample , and most of those data will be placed in publicly - accessible data archives after the surveys are completed . to date supporting data have been obtained on 20 telescopes , and when these surveys are completed they will comprise the most comprehensive multi - wavelength dataset ever assembled for a nearby galaxy sample . table 5 gives a compact summary of these ancillary and complementary datasets , including information on the field coverages , spatial resolutions , and approximate numbers of galaxies observed .
a more detailed galaxy - by - galaxy listing of ancillary and complementary observations will be maintained on the sings website and updated on a regular basis . in the remainder of this section
we briefly summarize each of these datasets .
broadband images in the visible and near - infrared provide an essential complement to the sirtf data , providing information on the local stellar spectral energy distributions at shorter wavelengths , and constraints on the dust extinction , stellar mass distributions , and stellar populations heating the dust .
emission - line images in h@xmath1 provide a local measurement of the stellar ionizing flux and sfr ( kennicutt 1983 ) , and when combined with the broadband data can constrain the ages and other properties of the star - forming regions . thanks to the generous allocation of observing time from noao , matched - field images in @xmath62 and h@xmath1
have been obtained for nearly all of the sings galaxies using the kpno 2.1 m telescope with cfim imager , and the ctio 1.5 m telescope with cassegrain ccd imager .
these imagers provide field coverage of 104 and 8 respectively , and for larger galaxies a series of positions were observed in order to provide full coverage to r @xmath63 r@xmath44 . for a handful of galaxies h@xmath1 images are being obtained with the 2k ccd imager on the steward observatory bok 2.3 m telescope .
all data have been obtained during photometric conditions , or under thin clouds with separate calibration exposures obtained in photometric weather .
two sets of near - infrared imaging ( @xmath64 ) of the sings galaxies are being collected .
a set of @xmath65 images will be extracted from the two - micron all sky survey ( 2mass ) .
these images will provide complete and relatively uniform quality near - infrared imaging for the entire sample , though the shallow depth of the data will limit applications for spatially - resolved studies ( cf .
jarrett et al .
deeper imaging is being obtained using the osiris imager on the ctio 1.5@xmath11 m telescope ( southern galaxies ) , and on the steward observatory 1.5@xmath11m(bigelow ) and 2.3@xmath11m(bok ) telescopes , using either the 256@xmath42256 infrared camera or the pisces wide - field camera ( mccarthy et al .
these cameras provide fields of view of 25 38 and 85 respectively .
since the observing time required to obtain high signal / noise at r@xmath44 would be prohibitive at these wavelengths , we have chosen instead to obtain high - quality imaging of the central regions of the galaxies , where extinction is most problematic at shorter wavelengths .
these images reach a typical depth of @xmath66 = 22 mag arcsec@xmath38 or @xmath67 = 21 mag arcsec@xmath38 . in order to provide more robust measurements of the extinction , massive star formation , and stellar mass distributions in the centers of the sings galaxies
, we are obtaining p@xmath1 emission - line and @xmath68-band imaging , using the nicmos camera on hst .
data for appoximately 30 sings galaxies are being obtained through an ongoing hst snapshot program ( go-9360 ) . since snapshot targets
are observed on an available opportunity basis , the exact number of galaxies to be observed ( and their identity ) is uncertain at this time .
in addition , archival observations are available for 19 galaxies , and we will process these data and deliver them as sings data products .
appropriate redshifted pa@xmath1 filters are only available for part of the sings sample , so coverage of the sample will be incomplete , but will include a wide enough range of properties to enable a wide range of scientific applications .
the images are being obtained with the nic3 camera , which provides a field size of 55 square . for typical exposure times and line / continuum ratios
the pa@xmath1 maps will reach a limiting surface brightness of @xmath69 ergs @xmath8 s@xmath19 arcsec@xmath38 over an area of 06 @xmath42 06 ( for @xmath70 ) .
the corresponding depth of the @xmath68-band broad band images will be 19.7 @xmath68 mag arcsec@xmath38 .
the p@xmath1 recombination line luminosity is directly tied to the ionizing luminosity in the same way as h@xmath1 , but due to its longer wavelength ( 1.87 ) it is much less susceptible to dust .
in addition the p@xmath1/h@xmath1 ratio provides a powerful extinction tracer in hii regions ( e.g. , bker et al .
1999 ; scoville et al .
2001 ; quillen & yukita 2001 ) , and it will provide a valuable independent constraint on the dust extinction and heating models in the central regions of these galaxies . as an aid to the interpretation of the sings 1040 and 5299 maps , our group is obtaining spectral drift - scans of the same regions in the 36007000 region , using steward observatory bok 2.3 m telescope for the northern galaxies and the ctio 1.5 m telescope for southern galaxies .
these data are obtained by trailing the galaxy image back and forth across the slit during an integration , so an integrated spectrum of a large rectangular strip in the sky can be obtained , while preserving the spectral resolution of the narrow - slit observation . for each galaxy
three one - dimensional integrated spectra are obtained .
one spectrum covers the same area as the irs and mips sed radial strips ( 4.4 ) , and a second is integrated over the central 20@xmath4220region , and is designed to provide matching spectrophotometry for the circumnuclear irs spectra ( 4.4 ) . a third
, narrow - slit pointed spectrum of the nucleus itself will also be provided .
the reduced , flux - calibrated spectra will have spectral resolution of @xmath1378 , and will provide information on the primary diagnostic nebular emission lines ( h@xmath1 , h@xmath71 , [ oii]@xmath163727 , [ oiii]@xmath164959,5007 , [ nii]@xmath166548,6583 , [ sii]@xmath166717,6731 ) , in some cases fainter lines such as [ oiii]@xmath164363 , and information on the shape and principle absorption lines in the stellar spectra . note
that since the spectra are taken in a drift - scanned mode there is no point - by - point spatial information .
optical spectra for most of the extra - nuclear hii regions in the sings sample ( table 3 ) are published in the literature , so new observations are not required .
a group from the universite de montreal and the laboratoire d astrophysique de marseille is obtaining h@xmath1 kinematics for the sings sample .
the observations will help to delineate the role of gas kinematics in regulating the sfr , and will provide a fundamental local reference sample for 3d kinematics studies of high - redshift galaxies .
all the data are being obtained with a new gaas photon counting system ( gach et al . 2002 ) .
the observations are obtained on the cfht 3.6 m , eso la silla 3.6 m , observatoire de haute provence ( ohp ) 1.93 m , and observatoire de mont megantic ( omm ) 1.6 m telescopes , with fields of view of 3 to 12 .
these data will complement the lower - resolution gas kinematic data being provided by the hi and co surveys ( below ) , by providing high spatial resolution data for the inner parts of galaxies .
these are crucial for constraining the dark matter density profiles of galaxies ( e.g. , blais - ouellette et al .
such a database will also make it possible to study how the parameters of the dark matter distribution vary as a function of morphological type , and test for influences on the star formation properties and histories of the galaxies .
imaging and spectra of the sings galaxies in the ultraviolet ( 13002800 ) provide an especially powerful complement to infrared data , because this wavelength region directly traces the photospheric emission of the dominant star - forming and dust - heating population in most galaxies .
the recently launched galaxy evolution explorer mission ( galex ) is designed to perform an all - sky survey in two bands over 13503000 , with angular resolution of @xmath63 5 ( milliard et al .
this is well matched to the resolution of sirtf , and will be adequate to quantify the integrated uv emission from the entire sample .
since the approval of the sings project the galex team also has tentatively planned much deeper imaging of each sings galaxy ( @xmath131h each at 13501800 and 18003000 ) .
these deeper images will enable point - by - point mapping of the seds and sfrs in the galaxies .
all of these data will be released into the galex public archive after a nominal proprietary period , and the images of the sings galaxies will be linked to our archive shortly thereafter .
images in this wavelength region ( 15002800 ) are also available for 27 of the galaxies from the ultraviolet imaging telescope ( uit ) and/or hst , and iue spectra of the central 10@xmath4220regions are available for 25 galaxies .
these data will be integrated into the sings database and included in the data archive as well .
maps of co and hi emission are important for tracing the cold , gaseous ism , and aperture synthesis maps enable a comparison with sirtf data at the highest resolution possible .
in addition , measures of radio continuum emission ( @xmath16 = 20 cm ) provide an alternate tracer of star formation as well as a means to examine the correlation of far - infrared and non - thermal radio continuum emission ( e.g. , condon 1992 ; marsh & helou 1995 ; sullivan et al .
2001 ) . for a core set of 20 spiral galaxies both co and hi aperture synthesis observations have been acquired as part of the bima survey of nearby galaxies ( bima song ) , and we will include the reduced maps in the sings data products ( regan et al . 2001 ; helfer et al .
the typical half - power beamwidth ( hpbw ) for these maps is 6 , and single - dish total - power measurements have been incorporated into the reductions to preserve extended emission .
further bima observations are planned for other northern sings galaxies as motivated by our sirtf observations
. published single - dish co measurements from a variety of sources are available for 36 of the sings galaxies ( see table 2 ) . in addition the five college radio astronomy observatory ( m. heyer , pi ) is performing an on - the - fly mapping survey of a broader subset of sings galaxies .
observations are being obtained using the fcrao 14 m telescope with the sequoua focal plane array ( 16 pixels for extra - galactic observations ) .
typical maps consist of 5@xmath425 to 8@xmath428 pointings , with a target sensitivity is 5 millikelvin in 5 mhz wide channels , chosen to match the sensitivity of the 12 m telescope observations from bima song .
we are assembling a database of 3045resolution hi maps for northern sings galaxies .
these data come from the dedicated westerbork ( wsrt ) hi survey , whisp ( swaters et al .
2002 ; swaters & balcells 2002 ) , as well as new and archival vla observations .
in addition the nearest sings galaxies ( d @xmath36 10 mpc ) are being observed in hi at high spatial and spectral resolution mapping as part of a vla large proposal ( f. walter , pi ) .
this will provide data for @xmath1335 galaxies of the sings sample at high spatial resolutions of 7 and velocity resolution 2.55kms@xmath19 .
these data will be used to investigate the small scale structure of the atomic ism and to obtain spatially resolved velocity dispersion maps for each galaxy .
the field of view of all hi observations is @xmath1330 , and nicely encompasses even the largest galaxies of the sings sample .
we are also assembling a database of 3045 resolution maps of the 20-cm radio continuum emission from sings galaxies .
these data come from the vla archives , as well as new observations of 40 sings galaxies , and seek to achieve an rms noise level half that of the nrao vla sky survey ( 0.20.3 mjybeam@xmath19 ; condon et al .
1998 ) at equal or higher resolution . in a parallel effort
a group led by robert braun ( nfra ) is using the westerbork synthesis radiotelescope ( wsrt ) to obtain deep 20 cm radio continuum and coarse - resolution hi line maps for 30 large ( d@xmath72 ) northern sings galaxies ( see table 4 ) .
these maps will have a resolution of approximately 12 @xmath42 12/@xmath73 .
the continuum data will include separate or combined 18 cm and 22 cm images reaching a limit of 15 @xmath11jybeam@xmath19 in each band ( including spectral index , polarization , and rotation measure maps ) . as can be seen from figures 1 and 8
, the thermal dust emission of galaxies usually extends to wavelengths well beyond 160 @xmath11 m , the longest wavelength for sirtf instruments .
this emission can represent a significant fraction of the total dust luminosity , and is critical for constraining the temperatures and masses of the cold dust component in the ism ( e.g. , siebenmorgen , krgel , & chini 1999 ; bendo et al .
2003 ) . to properly trace continuum dust emission from the far - infrared into the submillimeter , we have acquired data from the james clerck maxwell telescope ( jcmt ) archives maintained by the canadian astronomy data centre . at the time of this writing ,
the archive contain submillimetre common - user bolometer array ( holland et al .
1999 ) data for 24 sings galaxies .
only 17 galaxies are detected in at least the 850 waveband , although many of these galaxies are also detected at 450 .
for some galaxies , the data include stare photometry and jiggle maps of the inner 2.3 , whereas for other galaxies , the data include composite jiggle maps and scan maps that may extend beyond 10 apertures .
the data are being processed using the surf data reduction package ( jenness & lightfoot 1998 ) following standard data reduction procudures .
additional submillimeter observations are planned to acquire data for at least the brightest infrared sources in the sample .
data for 46 of the 75 sings galaxies are available from the infrared space observatory ( kessler et al .
1996 ) archives .
these data come from all four of the iso instruments , which we may use to supplement our existing data and check our flux calibration . the isocam ( cesarsky et al . 1996 ) and isophot ( lemke et al .
1996 ) broad band data may be useful for adding photometry to the observed spectral energy distributions of the sings galaxies , for checking the calibration of irs and mips data , and extending the sed coverage redward to 240 @xmath11 m .
isosws ( de graauw et al.1996 ) data provides spectra with higher spectral resolution than irs that can be used to closely examine the wavelength profiles of emission line features .
finally , isolws ( clegg et al .
1996 ) spectral data , which spans wavelengths longer than what irs can observe , can provide additional diagnostic spectral lines for modeling star formation and the interstellar medium . by the end of 2003 , 35 galaxies ( about 50% ) of the sings sample
will have been observed with the chandra x ray observatory ( with integration times ranging from 2 to 70ksec ) .
we anticipate that more sings targets will be added to the list in the years to come . as part of sings ,
processed x ray maps for these galaxies will be made available in various energy bands ( both photon and adaptively smoothed images ) .
chandra s arcsecond resolution will be essential for a detailed comparison of the x ray data with observations obtained at other wavelengths and for isolating point sources from the underlaying diffuse hot gas .
these x - ray point sources ( such as x ray binaries and supernova remnants ) provide important clues to the heating mechanisms of the ambient medium . in complementary fashion , the multi
wavelength sings data will be important to constrain the nature of the chandra point sources .
the chandra data will also allow us to derive physical properties ( such as tempertature , density , pressure ) of the hot phase of the ism which will be visible as diffuse , soft x ray emission .
the sings team is also collaborating on proposals to enlarge the subset of galaxies with high - quality chandra maps .
as with all legacy programs , the sirtf observations are completely non - proprietary .
this means that the pipeline - calibrated data , including calibrated single - pointing images ( and rudimentary mosaic images ) and individual spectra will be available to the public immediately .
these data will enable a wide range of scientific investigations in their own right , and provide invaluable samples of flight data for planning of future go observing programs .
a primary scientific output of the sings project will be added - value data products , including high - fidelity image mosaics , enhanced - resolution dithered images , spatially co - aligned multi - band images , one - dimensional spectral extractions , and full 3d spectroscopic data cubes . a typical image or spectroscopic observation of a sings galaxy
is built from 15815 individual sirtf datasets , so these data products will be the primary source of science - grade data for many users .
we also will deliver user versions of tools for extracting source spectral energy distributions from multi - wavelength sets of images and for building and analyzing spectral data cubes from sets of multiple irs pointings .
we will produce and deliver two sets of irac data products : 1 ) image mosaics in native irac resolution ; 2 ) enhanced mosaics .
the first set of images will possess the native resolution of the instrument , which is limited by the pixel size of the irac detectors ( 12 ) .
since this pixel size undersamples the telescope s point spread function at all wavelengths , we plan to construct a second set of mosaic images that are processed using the drizzle technique ( fruchter & hook 2002 ) to reconstruct the point spread function , and thus provide higher spatial resolution images .
the processing and production of the mosaic images will also include a number of other checks and enhancements to the basic calibrated data .
cross - correlation of the individual pointings will be used to derive spatial offsets between subexposures , and check the pointing reconstruction given in the image headers .
we will also check for and if needed remove other instrumental signatures such as detector bias level drifts , residual cosmic rays , and residual hot pixels .
asteroids will be detected by combining the time - separated exposures that are being designed for this purpose .
the mips data products will include 24 , 70 , and 160 mosaicked images and a sed strip spectral cube for each of the 75 sings galaxies .
the initial processing will proceed in similar fashion to irac , with a combination of manual checks on data quality followed by a carefully supervised processing of individual observations to create mosaicked images .
the mips image data products will be delivered in two versions .
the first version will encompass the entire area observed at each wavelength . due to the placement of the three focal plane arrays and their different sizes , this will result in different fields - of - view at each wavelength for both photometry and scan mode observations .
the second version will have images which are background - subtracted ( constant value ) , share a common field - of - view , and have been regridded to a common pixel scale .
the mips sed data will be processed in the same way , before being combined with the cubism package ( 6.3.2 ) to create spectral data cubes from individual mosaicked sed pointings .
the resulting data products will be delivered in two versions as well .
the first will consist of two spectral cubes , one for the object ( without background removal ) and one for the nearby background observations
. the second version will be a single spectral cube with the nearby background observations subtracted from the object observations .
a program for automated mosaicking of mips images ( named mips_enhancer ) is being developed by the mips instrument team in collaboration with the sings team .
the goal is to use the redundancy inherent in the mips observational strategy to remove residual instrumental signatures in the 70 and 160 data .
the program will be continually improved as knowledge of the detector instrumental signatures increases , and a final delivery of mosaicked images and spectra that incorporate our accumulated knowledge to date will be made at the end of the project .
the sings irs data products consist of three separate types for each mapped region : 1 ) full , 3d spectral data cubes ; 2 ) selected line and continuum spectral maps ; and 3 ) representative one - dimensional extracted spectra . tools for creating and visualizing the spectral cubes will also be available , and these are described below ( 6.3.2 ) .
the fundamental data product generated from the irs spectral mapping observations is a spectral cube , created by combining spatially overlapping data from different slit pointings into a three dimensional data cube with coordinate axes @xmath74 .
the other irs data products are derived directly from the spectral cube .
a single spectral cube will be created for each irs sub - slit ( 6 in total ) used in all nuclear , extra - nuclear and radial strip observations .
the cubes are constructed using a flux - conservative algorithm somewhat related to drizzle ( fruchter & hook 2002 ) , but operating in three dimensions on collections of 2d spectral images .
spatial and spectral smoothing , and other error - correlating operations are avoided by mapping pixels from the raw spectra directly to the spectral cube .
individual single - wavelength spatial samples , consisting of tilted `` pseudo - rectangles '' approximately as long as the slit , and with widths equal to the wavelength sampling interval , are tiled on the detector , and clipped against its pixel grid .
the resulting partial pixels are rotated to match the `` sky grid '' of the final cube , both to remove any intrinsic instrumental spectral line tilt , and to reference all observations to a common position angle of the slit on the sky .
samples collected from different spectral orders which overlap in wavelength , and from adjacent spectrograms within the mapping observation set are combined together in the resultant cube .
the entire set of rotated partial pixels is clipped again against the output sky grid , which is constructed with an adjustable pixel spacing defaulting to the native plate scale of the irs module in use ( none of the irs modules is significantly undersampled ) .
the clipped area for each pixel portion is retained , and in this way an area - conserving mapping of spectrogram pixels to cube pixels is built up , and used to construct the final cube .
area / error weighting is the default method for combining the contributing pixels , but trimmed weighted means are also possible .
spectral maps in [ neii ] ( 12.8@xmath11 m , @xmath75=22ev ) , [ neiii ] ( 15.5@xmath11 m , @xmath75=41ev ) , and [ siii ] ( 34.8@xmath11 m , @xmath75=8ev ) line features will be produced from all the sings nuclei and targeted hii region cubes .
these maps will be generated from the spectral cubes based on simple fits to the underlying continuum .
the continuum - subtracted maps , along with the derived continuum maps used in the subtraction will be included .
in addition to the ne and si atomic line maps , we will produce maps in the combined [ 7.7@xmath11 m + 8.6@xmath11 m ] , and 11.3@xmath11 m pah features for all the sings nuclei and targeted hii regions .
since these broad dust features dominate the spectrum of star - forming galaxies at these wavelengths , we will not perform continuum subtraction . instead
, we will extract a set of line - free continuum maps that will provide information on the underlying continuum , as well as providing additional information on the galaxy sed at wavelengths that fall between the irac and mips filters ( 1440 ) . by selecting spatial regions in the maps or cubes
we can create 1d spectral extractions for a given irs sub - slit .
we will produce 1d spectra for all sings nuclei and mapped hii regions . these source
" spectra will be extracted from a suitable region ( matched to the psf ) , centered on the map peak at either 12 or 25@xmath11 m , depending upon the irs module used to create the map .
the surrounding areas in the map will be used to create a local background " 1d spectrum , which will suitably scaled and subtracted from the source spectrum .
both the background - subtracted source spectrum and the background spectrum itself will be included in the data products .
most of the sings data products will be in the form of image mosaics and extracted spectra which can be readily analyzed by the generic sirtf data analysis programs or by standard image analysis packages such as iraf or idl .
however our team is developing two software packages that will be specifically tailored to analysis of the sings data .
one of these tools ( sedtiap ) will enable the batch - mode generation of psf - matched seds from a set of individual multi - wavelength images .
the cubism tool will generate 3d spectral data cubes from irs mapping observations , and provide a variety of user software for analyzing the spectral data cubes .
the sedtiap program is an idl - based tool that rotates , rescales , interpolates / convolves the input image data to a common astrometric coordinate system and resolution , using the user - selected psf from a set of available choices and assuming bi - linear interpolation with flux conservation to resample the images .
the co - aligned images are displayed for the user to check alignment and a data - cube containing the processed images is saved to disk .
apertures for both sources and background are specified interactively by the user , and the total and background - subtracted fluxes for all the images in the data - cube are saved in a file ( the sed file ) .
spectral cubes are constructed from the irs spectral mapping data sets using a custom software package , cubism .
the cubes themselves and the algorithms used to create them are discussed in [ sec : spectral - cubes ] .
cubism is written in idl , and consists of four main components : i ) the cube - builder backend ( see [ sec : spectral - cubes ] ) ; ii ) cubeproject , a graphical tool for organizing and querying cube inputs and driving the cube creation ; iii ) cubeview , a visualizer for both 2d spectrograms and spectral cubes , facilitating 1d extractions ; iv ) cubespec , for visualizing 1d extracted cube spectra , and generating stacked maps .
cubeproject provides the interface from which all cube - building operations are performed .
it encapsulates individual data records into `` projects '' , each of which includes all the calibration information and data required to build a single cube for one irs sub - slit .
individual data records can be added , removed , disabled , queried for header information , and sorted in a variety of ways .
visual feedback of the pixel clipping process is provided , and the individual spectrograms or the final cube can be sent to cubeview for display .
cubeview permits both individual spectral data records and the full cube to be displayed and analyzed .
a suite of traditional viewer tools ( zooming , colormap and histogram scaling , box statistics , photometry , etc . ) is available .
in addition , various specialized tools are available for reporting wavelength and order ( for spectrograms ) or celestial position ( for spectral cubes ) within the image , selecting and setting the apertures which delineate the order data to be included in the cube build , navigating through wavelength planes , and extracting spectra from selected regions . in the latter case
, the extracted 1d spectrum is displayed in a related tool , cubespec , which can be used to visit individual planes of the cube and to create line or feature maps .
arbitrary peak and continuum wavelength regions can be defined on the spectrum , and the resultant maps are displayed in cubeview , where they can be saved or exported . stored
region sets for important lines and features in the wavelength range are available , including 8 and 24@xmath11 m sets corresponding to the appropriate irac and mips bandpasses , to enable easy cross - calibration between instruments .
the continuum can also be fit and removed using the defined background regions .
as discussed in 5 , the scientific value of the sirtf data will be enhanced substantially by the addition of images and spectra of the galaxies at other wavelengths .
some of these data ( e.g. , visible and near - infrared imaging , visible spectra ) are being obtained by the sings team and will be delivered to the nasa / ipac infrared space archive ( irsa ) as formal ancillary data products . they will be fully reduced and will include headers with full world coordinate system ( wcs ) information , flux calibration information , and when relevant comments on the photometric and/or astrometric accuracy of the images .
data from other sources that are re - reduced by our team ( e.g. , submillimeter , x - ray images ) will also carry comments or documentation on the reduction procedures and any qualifications on the data quality .
complementary data from other groups ( table 5 ) will be posted either in the sings archive or on independent websites , according to the wishes of the groups involved .
information and links to these sites will be provided on the main sings website ( below ) .
all public access to sings data will be made through the sirtf legacy archive at irsa .
data releases and access will be administered by the sirtf science center , and prospective users should follow announcements from ssc to obtain more updated information .
in addition the sings team maintains its own public website .
the site contains detailed information on the project and galaxy sample , and will be regularly updated to provide information on observations and data products .
scheduling of the sirtf observations and the subsequent data releases is subject to the launch date , and scheduling of the various gto , legacy , and go observations thereafter .
however the graph in figure 9 provides an approximate projection of the expected data flow , as a function of month after launch .
we expect approximately 5 galaxies per month to be observed on average , beginning approximately 34 months after the launch of sirtf .
the first of these data will be released to the public with the opening of the sirtf archive , approximately 8 months after launch , and after that time the basic calibrated data and other pipeline observations will be posted in the archive shortly after the observations are taken ( as indicated by the leftmost dashed line in figure 9 .
a small fraction of the spectroscopic observations ( galaxies with questionable detections , ir - selected extranuclear targets ) will not be scheduled until irac and mips images for those galaxies are in hand , and this is indicated by the turnover in the dashed line at the top of the figure . in order to expedite the release of science - grade products to the user community we have defined two levels of product for each instrument , as summarized in table 6 .
the rapid release " products consist of processed and mosaicked images for irac and mips , and 1d spectral extractions and line maps for irs .
each product will incorporate full checking and validation of the pipeline data by members of the sings team , calibration and astrometric checks and refinements , combination of individual observations into mosaic images or spectral maps , and corrections for instrumental signatures on a best - effort basis , using the knowledge we have acquired at the time .
approximately 12 months later the sings team will deliver a second generation set of enhanced " data products ( table 5 ) .
these will include improved resolution dithered irac mosaic images , mips images which take into account our accumulated experience with removing detector transient features , full spectral data cubes and the associated cubism and cubeview tools , and if necessary improved calibrations for all relevant data products .
our current projection for these deliveries is shown by the rightmost dot - dashed line in figure 8 . in order to maximize the scientific value of the sirtf data products for a given galaxy
, releases will include the relevant ancillary and complementary data for that galaxy ( as available ) , to provide a coherent multi - wavelength dataset .
it is anticipated that all of the data described in this paper will be in the public archive by the end of the project ( approximately 3 years after the sirtf launch ) .
we are pleased to acknowledge the support of the staff at ssc and sirtf , with special thanks to michael bicay , tony marston , and our ssc liaison scientist , nancy silbermann .
likewise we thank the director s office and staff at kpno and ctio for their generous award of observing time for ancillary observations , and their capable support during the observing runs .
we also extend special thanks to irs pi jim houck and his team for cooperating on observations of common targets , and to the irs smart team , as well as the rest of the irs data products interest group , for their cooperation on the development of irs reduction software .
we are also indebted to mips pi george rieke and his team for their close cooperation and collaboration in the development of mapping and data processing software for observations of extended objects .
we are also very grateful to john moustakas and eric murphy for assisting in the ancillary observations and observing planning .
finally we thank philippe amram , robert braun , claude carignan , mark heyer , bob joseph , thijs van der hulst for leading projects to obtain additional ancillary data for the sings galaxies .
support for this work , part of the space infrared telescope facility ( sirtf ) legacy science program , was provided by nasa through an award issued by the jet propulsion laboratory , california institute of technology under nasa contract 1407 .
jenness , t. , & lightfoot , j. f. 1998 , in albrecht , r. , hook , r. n. , bushouse , h. a. eds , asp conf .
145 , astronomical data analysis software and systems vii .
, san francisco , p. 216
ll hubble type & e irr @xmath76 & @xmath7712.5 @xmath7723.5 l@xmath78 & @xmath79 @xmath80 l@xmath81 ( 0.000310 l@xmath82 ) l(ir ) & @xmath3610@xmath83 @xmath84 l@xmath81 l(ir)/l@xmath85 & @xmath360.02 42 f(60)/f(100 ) & 0.16 1.2 o / h ( 0.4 @xmath86 ) & 0.05 3 ( o / h)@xmath81 m@xmath87/m@xmath88 & @xmath360.001 5 sfr & 0 15 m@xmath21/yr sfr/@xmath89 & @xmath3610@xmath90 10@xmath91 m@xmath81/yr/@xmath92 lrrrcrrr ngc 628 h 292 & 01:36:45.10 & 15:47:51 & 4.0 & 9.1 & 2.0 & 40.4 & 45 ngc 628 h 572 & 01:36:37.50 & 15:45:12 & 6.8 & 9.0 & 0.9 & 39.5 & 6 ngc 628 h 627 & 01:36:38.80 & 15:44:25 & 8.9 & 8.8 & 0.9 & 40.2 & 30 ngc 628 h 13 & 01:36:35.50 & 15:50:11 & 11.7 & 8.6 & 1.5 & 40.1 & 20 & & & & & & & ngc 2403 hk 361 & 07:36:46.55 & 65:36:54 & 1.0 & 8.8 & 0.4 & 39.5 & 60 ngc 2403 hk 270 & 07:36:53.50 & 65:36:40 & 1.1 & 8.7 & 0.3 & 39.4 & 45 ngc 2403 hk 128 & 07:37:08.19 & 65:36:33 & 2.8 & 8.6 & 0.8 & 39.9 & 150 ngc 2403 hk 65 & 07:37:19.48 & 65:33:57 & 3.5 & 8.6 & 1.8 & 39.6 & 70 ngc 2403 hk 542 & 07:36:21.05 & 65:36:55 & 3.7 & 8.6 & 0.5 & 39.6 & 80 ngc 2403 vs 9 & 07:36:29.94 & 65:33:43 & 6.2 & 8.1 & & 39.0 & 20 & & & & & & & ho ii hsk 45 & 08:19:13.30 & 70:43:08 & & 8.5 & & 39.1 & 22 ho ii hsk 67 & 08:19:27.00 & 70:41:59 & & 8.6 & & 38.5 & 5 ho ii hsk 70 & 08:19:28.80 & 70:42:21 & & & & 38.6 & 8 ho ii hsk 7 & 08:18:50.10 & 70:44:48 & & & & 38.5 & 6 & & & & & & & m81 hk 230 & 09:56:00.44 & 69:04:02 & 4.1 & 9.2 & 0.4 & 38.7 & 10 m81 hk 343 & 09:55:40.67 & 68:59:45 & 4.7 & 8.9 & 0.0 & 38.6 & 8 m81 hk 453 & 09:55:24.43 & 69:08:15 & 4.8 & 8.9 & 1.0 & 39.0 & 20 m81 hk 268 & 09:55:53.17 & 68:59:04 & 5.1 & 8.8 & 0.7 & 39.3 & 40 m81 hk 652 & 09:54:56.63 & 69:08:47 & 5.6 & 8.8 & 0.7 & 39.0 & 20 m81 hk 741 & 09:54:42.26 & 69:03:36 & 7.7 & 8.7 & 0.5 & 39.1 & 25 m81 munch 1 & 09:56:17.42 & 68:49:50 & 15.0 & 8.1 & 0.2 & 38.0 & 23 & & & & & & & ngc 4559 zkh 20 & 12:35:56.50 & 27:57:40 & 1.4 & 8.9 & 1.1 & 39.5 & 8ngc 4559 zkh 18 & 12:36:00.80 & 27:56:22 & 4.1 & 8.8 & 0.7 & 39.9 & 18 ngc 4559 zkh 17 & 12:36:02.40 & 27:56:45 & 4.7 & 8.8 & 1.1 & 39.7 & 10 ngc 4559 zkh 4 & 12:35:52.10 & 27:59:10 & 12.3 & 8.6 & 0.4 & 39.3 & 5 & & & & & & & ngc 4736 hk 52 & 12:50:49.64 & 41:07:23 & 0.9 & 9.1 & 0.5 & 39.0 & 10 ngc 4736 hk 53 & 12:50:49.64 & 41:07:34 & 1.0 & 9.0 & 1.0 & 39.0 & 10 ngc 4736 hk 8/9 & 12:50:56.27 & 41:07:20 & 1.0 & 9.0 & 0.5 & 39.3 & 20 & & & & & & & m51 ccm 107 & 13:29:53.10 & 47:12:40 & 2.5 & 9.4 & 1.7 & 39.9 & 35 m51 ccm 72 & 13:29:44.10 & 47:10:21 & 5.0 & 9.3 & 1.9 & 40.3 & 70 m51 ccm 71 & 13:29:44.60 & 47:09:55 & 5.3 & 9.3 & 0.9 & 39.8 & 23 m51 ccm 1 & 13:29:56.20 & 47:14:07 & 5.8 & 9.2 & 2.6 & 40.1 & 50 m51 ccm 10 & 13:29:59.60 & 47:14:01 & 6.4 & 9.2 & 2.2 & 39.8 & 23 m51 cm 71a & 13:29:39.50 & 47:08:35 & 9.5 & 9.1 & 2.7 & 40.2 & 59 & & & & & & & ngc 6822 hub v & 19:44:52.85 & @xmath7714:43:11 & & 8.1 & & 38.3 & 33 ngc 6822 hub x & 19:45:05.24 & @xmath7714:43:13 & & 8.1 & & 38.2 & 75 ngc 6822 hub i & 19:44:31.64 & @xmath7714:42:01 & & 8.1 & & 37.9 & 120 ngc 6822 hub iii & 19:44:34.05 & @xmath7714:42:22 & & 8.1 & & 37.5 & 10 & & & & & & & ngc 6946 h 4 & 20:35:16.68 & 60:10:57 & 6.0 & 9.1 & 2.2 & 39.8 & 40 ngc 6946 hk 3 & 20:35:25.12 & 60:10:03 & & 9.0 & 1.5 & 40.0 & 75 ngc 6946 hk 288 & 20:34:52.29 & 60:12:41 & & 9.0 & 1.5 : & 39.7 & 35 ngc 6946 h 40 & 20:34:19.48 & 60:10:09 & 7.7 & 8.9 & 1.2 & 39.5 & 20 ngc 6946 h 28 & 20:34:39.03 & 60:13:35 & 8.6 & 8.9 & 1.2 & 39.5 & 20 lrrllrrrrrrrrrrl ngc 0024 & 00:09:56.7 & -24:57:44 & sac & & 554 & 8.2 & [email protected] & @xmath7718.4 & 222 .
& 0.20 & 0.35 & 9.07 & & & 19 -8 + 7 ngc 0337 & 00:59:50.3 & -07:34:44 & sbd & & 1650 & 24.7 & [email protected] & @xmath7720.3 & 264 .
& 1.20 & 0.49 & & & 4.3 & 61 -23 + 21 ngc 0584 & 01:31:20.7 & -06:52:05 & e4 & & 1854 & 27.6 & [email protected] & @xmath7722.2 & & & & & & & 52 -7 ngc 0628 & 01:36:41.8 & + 15:47:00 & sac & & 657 & 11.4 & [email protected] & @xmath7720.9 & 74 .
& 0.72 & 0.31 & 10.11 & 9.49 & 4.0 & 17 -4 ngc 0855 & 02:14:03.6 & + 27:52:38 & e & & 610 & 9.6 & [email protected] & @xmath7717.7 & & 0.23 & 0.48 & & & & ngc 0925 & 02:27:16.9 & + 33:34:45 & sabd & h & 553 & 10.1 & [email protected] & @xmath7720.6 & 224 . & 0.18 & 0.29 & 9.79 & @xmath369.03 & 2.4 & 17 -1 ngc 1097 & 02:46:19.0 & -30:16:30 & sbb & l & 1275 & 16.9 & [email protected] & @xmath7722.4 & 402 .
& 0.75 & 0.40 & 10.03 & & 4 - 8 : & 51 -3 + 1 ngc 1266 & 03:16:00.7 & -02:25:38 & sb0 & l & 2194 & 31.3 & [email protected] & @xmath7721.1 & & 5.49 & 0.76 & & & & ngc 1291 & 03:17:18.6 & -41:06:29 & sba & & 839 & 9.7 & [email protected] & @xmath7722.0 & 84 . & 0.02 & 0.17 & 9.19 & & 0.4 & 53 -9 + 7 ngc 1316 & 03:22:41.7 & -37:12:30 & sab0 & l & 1760 & 26.3 & [email protected] & @xmath7723.5 & & 0.04 & 0.41 & @xmath368.87 & & & 51 -1 ngc 1377 & 03:36:39.1 & -20:54:08 & s0 & h & 1792 & 24.4 & [email protected] & @xmath7719.6 & & 2.04 & 1.23 & & & & ngc 1404 & 03:38:51.9 & -35:35:37 & e1 & & 1947 & 25.1 & [email protected] & @xmath7722.9 & & & & & & & 51 -1 ngc 1482 & 03:54:39.3 & -20:30:09 & sa0 & & 1655 & 22.0 & [email protected] & @xmath7720.5 & & 4.65 & 0.77 & @xmath368.88 & 9.47 & 1.5 & 51 -4 ngc 1512 & 04:03:54.3 & -43:20:56 & sbab & sb & 896 & 10.4 & [email protected] & @xmath7719.9 & 271 .
& 0.24 & 0.29 & 9.77 & & & 53 -7 ngc 1566 & 04:20:00.4 & -54:56:16 & sabbc & sy & 1496 & 18.0 & [email protected] & @xmath7721.9 & 233 . & 0.49 & 0.32 & 10.03 & & & 53 -1 ngc 1705 & 04:54:13.5 & -53:21:40 & am & sb & 628 & 5.8 & [email protected] & @xmath7716.7 & & 0.29 & 0.48 & & & 0.03 & 53 + 1 ngc 2403 & 07:36:51.4 & + 65:36:09 & sabcd & h & 131 & 3.5 & [email protected] & @xmath7719.7 & 257 . &
0.29 & 0.35 & 9.73 & 7.86 & 1.3 & 14 -10 ho ii & 08:19:05.0 & + 70:43:12 & i m & & 157 & 3.5 & [email protected] & @xmath7717.1 & 73 . &
0.06 & 0.44 & & & 0.12 & 14 -10 m81 dwa & 08:23:56.0 & + 71:01:45 & i ? & & 113 & 3.5 & [email protected] & & 38 .
& & & @xmath367.06 & 6.33 & @xmath360.001 & 14 -10 ddo 053 & 08:34:07.2 & + 66:10:54 & i m & & 19 & 3.5 & [email protected] & @xmath7713.6 & & 0.48 & 0.29 & & & 0.005 & 14 -10 ngc 2798 & 09:17:22.9 & + 41:59:59 & sba & sb & 1726 & 24.7 & [email protected] & @xmath7719.6 & 316 . & 4.71 & 0.70 & 9.29 & 9.47 & 2.0 & 21 -16 ngc 2841 & 09:22:02.6 & + 50:58:35 & sab & l / sy & 638 & 9.8 & [email protected] & @xmath7720.7 & 611 . & 0.11 & 0.18 & 9.53 & 9.30 & 0.2 & 15 + 10 ngc 2915 & 09:26:11.5 & -76:37:35 & i0 & sb & 468 & 2.7 & [email protected] & @xmath7715.1 & 157 . & 0.31 & 1.17 & 8.25 & & 0.04 & 14 + 20 ho i & 09:40:32.3 & + 71:10:56 & iabm & & 143 & 3.5 & [email protected] & @xmath7713.2 & 44 . & & & 8.15 & & 0.004 & 14 -10 ngc 2976 & 09:47:15.4 & + 67:54:59 & sac & h & 3 & 3.5 & [email protected] & @xmath7717.6 & & 0.49 & 0.37 & 8.26 & 7.91 & 0.2 & 14 -10 ngc 3049 & 09:54:49.6 & + 09:16:18 & sbab & sb & 1494 & 19.6 & [email protected] & @xmath7718.7 & 213 . & 0.64 & 0.67 & 9.10 & & & ngc 3031 & 09:55:33.2 & + 69:03:55 & saab & l & @xmath7734 & 3.5 & [email protected] & @xmath7721.2 & 446 . & 0.08 & 0.26 & 8.88 & & 1.1 & 14 -10 ngc 3034 & 09:55:52.2 & + 69:40:47 & io & sb & 203 & 3.5 & [email protected] & @xmath7717.9 & & 42.35 & 0.97 & 8.85 & 9.39 & 6.0 : & 14 -10 ho ix & 09:57:32.0 & + 69:02:45 & i m & & 46 & 3.5 & [email protected] & @xmath7713.6 & & & & 7.55 & & 0.001 & m81 dwb & 10:05:30.6 & + 70:21:52 & i m & & 350 & 3.5 & [email protected] & @xmath7712.5 & 69 . & & & 7.08 & & 0.004 & 14 -10 ngc 3190 & 10:18:05.6 & + 21:49:55 & saap & l & 1271 & 17.4 & [email protected] & @xmath7720.7 & 601 . &
0.30 & 0.32 & 8.65 & & @xmath360.1 & 21 -6 ngc 3184 & 10:18:17.0 & + 41:25:28 & sabcd & h & 592 & 8.6 & [email protected] & @xmath7719.0 & 142 . & 0.99 & 0.31 & 9.32 & 8.96 & 1.2 & 15 + 7 ngc 3198 & 10:19:54.9 & + 45:32:59 & sbc & & 663 & 9.8 & [email protected] & @xmath7720.2 & 318 . & 0.19 & 0.40 & 9.74 & & 0.85 & 15 + 7 ic 2574 & 10:28:21.2 & + 68:24:43 & sabm & & 57 & 3.5 & [email protected] & @xmath7717.7 & 123 . & 0.11 & 0.23 & 9.21 & & 0.10 & 14 -10 ngc 3265 & 10:31:06.8 & + 28:47:47 & e & & 1421 & 20.0 & [email protected] & @xmath7717.7 & & 1.30 & 0.65 & 8.24 & & & mrk 33 & 10:32:31.9 & + 54:24:03 & i m & sb & 1461 & 21.7 & [email protected] & @xmath7718.4 & 181 . &
2.29 & 0.88 & 8.77 & & 1.5 & 13 + 1 ngc 3351 & 10:43:57.7 & + 11:42:13 & sbb & sb & 778 & 9.3 & [email protected] & @xmath7720.4 & 288 . & 0.54 & 0.51 & 9.10 & 8.82 & 1.2 & 15 -1 ngc 3521 & 11:05:48.6 & -00:02:09 & sabbc & l & 805 & 9.0 & [email protected] & @xmath7721.0 & 466 . & 0.81 & 0.38 & 9.67 & 9.64 & 1.7 & 15 -0 + 1 ngc 3621 & 11:18:16.3 & -32:48:45 & sad & & 727 & 6.2 & [email protected] & @xmath7719.4 & 290 . & 1.10 & 0.33 & 9.88 & & 5.1 & 15 -0 ngc 3627 & 11:20:15.0 & + 12:59:30 & sabb & sy2 & 727 & 8.9 & [email protected] & @xmath7720.8 & 378 . & 1.21 & 0.47 & 8.92 & 9.61 & 6.9 & 15 -2 + 1 ngc 3773 & 11:38:13.0 & + 12:06:43 & sa0 & & 987 & 12.9 & [email protected] & @xmath7717.5 & 191 . &
0.40 & 0.82 & 7.99 & & & 21 -4 + 1 ngc 3938 & 11:52:49.4 & + 44:07:15 & sac & & 809 & 12.2 & [email protected] & @xmath7720.1 & 112 . & 0.46 & 0.33 & 9.57 & 9.46 & 1.2 & 12 -1 ngc 4125 & 12:08:06.0 & + 65:10:27 & e6p & & 1356 & 21.4 & [email protected] & @xmath7721.6 & & 0.03 & 0.39 & & & & 12 + 5 -1 ngc 4236 & 12:16:42.1 & + 69:27:45 & sbdm & & 0 & 3.5 & [email protected] & @xmath7718.1 & 176 . & 0.09 & 0.40 & 9.23 & @xmath368.14 & 0.3 & 14 + 10 ngc 4254 & 12:18:49.6 & + 14:24:59 & sac & & 2407 & 20.0 & [email protected] & @xmath7721.6 & 272 . & 1.02 & 0.37 & 9.86 & 10.12 & 11.0 & 11 -1 ngc 4321 & 12:22:54.9 & + 15:49:21 & sabbc & l & 1571 & 20.0 & [email protected] & @xmath7722.1 & 283 . &
0.73 & 0.37 & 9.67 & 10.17 & 5.5 & 11 -1 lrrllrrrrrrrrrrl ngc 4450 & 12:28:29.6 & + 17:05:06 & saab & l & 1954 & 20.0 & [email protected] & @xmath7721.4 & 290 . &
0.07 & 0.19 & 8.61 & 9.30 & 0.5 & 11 -1 ngc 4536 & 12:34:27.1 & + 02:11:16 & sabbc & h & 1808 & 25.0 & [email protected] & @xmath7720.8 & 337 .
& 2.33 & 0.64 & 9.71 & 9.71 & 3.7 & 11 -4 + 1 ngc 4552 & 12:35:39.9 & + 12:33:22 & e0 & l & 340 & 20.0 & [email protected] & @xmath7720.8 & & & & & & & 11 -1 ngc 4559 & 12:35:57.7 & + 27:57:35 & sabcd & h & 816 & 11.6 & [email protected] & @xmath7721.0 & 251 . & 0.17 & 0.41 & 10.05 & & & 14 -1 ngc 4569 & 12:36:49.8 & + 13:09:46 & sabab & l / sy & @xmath77235 & 20.0 & [email protected] & @xmath7722.0 & 360 .
& 0.22 & 0.37 & 8.80 & 9.82 & 1.9 & 11 -1 ngc 4579 & 12:37:43.6 & + 11:49:05 & sabb & l / sy & 1519 & 20.0 & [email protected] & @xmath7721.8 & 390 . & 0.17 & 0.28 & 8.91 & 9.60 & 2.0 & 11 -1 ngc 4594 & 12:39:59.4 & -11:37:23 & saa & l / sy2 & 1091 & 13.7 & [email protected] & @xmath7721.5 & 762 . & 0.16 & 0.19 & 8.77 & @xmath369.36 & 0.1 & 11 -14 + 10 ngc 4625 & 12:41:52.7 & + 41:16:25 & sabmp & & 609 & 9.5 & [email protected] & @xmath7717.5 & 86 . & 0.37 & 0.34 & 9.02 & & & 14 -4 ngc 4631 & 12:42:08.0 & + 32:32:26 & sbd & & 606 & 9.0 & [email protected] & @xmath7720.6 & 320 . & 1.26 & 0.40 & 10.09 & 9.19 & 3.3 & 14 -6 ngc 4725 & 12:50:26.6 & + 25:30:03 & sabab & sy2 & 1206 & 17.1 & [email protected] & @xmath7722.0 & 410 . & 0.09 & 0.20 & 9.87 & 9.80 & & 14 -2 + 1 ngc 4736 & 12:50:53.0 & + 41:07:14 & saab & l & 308 & 5.3 & [email protected] & @xmath7719.9 & 241 . &
0.87 & 0.57 & 8.94 & 8.90 & 2.1 & 14 -7 ddo 154 & 12:54:05.2 & + 27:08:59 & ibm & & 376 & 5.4 & [email protected] & @xmath7715.1 & 103 . & & & 8.99 & & 0.0015 & 14 + 3 ngc 4826 & 12:56:43.7 & + 21:40:52 & saab & sy2 & 408 & 5.6 & [email protected] & @xmath7720.3 & 311 . & 0.27 & 0.48 & 8.49 & 8.87 & 0.3 & 14 + 3 ddo 165 & 13:06:24.8 & + 67:42:25 & i m & & 37 & 3.5 & [email protected] & @xmath7715.3 & 68 .
& & & 8.02 & & 0.002 & 14 + 10 ngc 5033 & 13:13:27.5 & + 36:35:38 & sac & sy2 & 875 & 13.3 & [email protected] & @xmath7720.9 & 446 . & 0.48 & 0.34 & 9.97 & 9.50 & 2.1 & 43 -1 ngc 5055 & 13:15:49.3 & + 42:01:45 & sabc &
h / l & 504 & 8.2 & [email protected] & @xmath7719.0 & 405 .
& 4.56 & 0.25 & 9.88 & 9.62 & 2.3 & 14 -5 ngc 5194 & 13:29:52.7 & + 47:11:43 & sabbc & h / sy2 & 463 & 8.2 & [email protected] & @xmath7721.4 & 195 . & 0.60 & 0.35 & 9.60 & 9.83 & 5.4 & 14 -5 ngc 5195 & 13:29:58.7 & + 47:16:05 & sb0p & l & 552 & 8.2 & [email protected] & @xmath7720.0 & & 0.29 & 0.53 & & 8.25 & @xmath360.1 & 14 -5 ngc 5398 & 14:01:21.3 & -33:03:47 & sbdm & h & 1216 & 15.0 & [email protected] & @xmath7718.9 & 137 . & 0.44 & 0.58 & 9.11 & & & 16 -5 ngc 5408 & 14:03:20.9 & -41:22:40 & ibm & & 509 & 4.5 & [email protected] & @xmath7716.1 & 114 .
& 0.74 & 0.89 & 8.45 & & 0.18 & 14 -15 ngc 5474 & 14:05:01.6 & + 53:39:44 & sacd & h & 273 & 6.9 & [email protected] & @xmath7718.4 & 61 . & 0.11 & 0.28 & 9.10 & & 0.2 & 14 -9 ngc 5713 & 14:40:11.5 & -00:17:21 & sabbcp & & 1883 & 26.6 & [email protected] & @xmath7720.9 & 209 . & 1.71 & 0.57 & 9.93 & 9.72 & & 41 -2 + 1 ngc 5866 & 15:06:29.5 & + 55:45:48 & s0 & & 692 & 12.5 & [email protected] & @xmath7719.9 & & 0.51 & 0.29 & @xmath368.28 & 8.63 & @xmath360.1 & 44 -1 ic 4710 & 18:28:38.0 & -66:58:56 & sbm & sb & 741 & 8.5 & [email protected] & @xmath7718.3 & 51 . & 0.20 & 0.38 & 8.69 & & 1.3 & 19 -1 ngc 6822 & 19:44:56.6 & -14:47:21 & ibm & & @xmath7757 & 0.6 & [email protected] & @xmath7713.8 & 81 .
& 2.50 & 0.50 & 8.26 & & 0.024 & 14 -12 ngc 6946 & 20:34:52.3 & + 60:09:14 & sabcd & h & 48 & 5.5 & [email protected] & @xmath7721.3 & 242 .
& 0.39 & 0.40 & 9.79 & 9.61 & 2.2 & 14 -0 ngc 7331 & 22:37:04.1 & + 34:24:56 & sab & l & 816 & 15.7 & [email protected] & @xmath7721.8 & 530 .
& 1.02 & 0.30 & 10.01 & 10.05 & 4.2 & 65 -1 ngc 7552 & 23:16:11.0 & -42:34:59 & sac & sb / l & 1585 & 22.3 & [email protected] & @xmath7721.7 & 280 .
& 3.20 & 0.71 & 9.68 & & 7.0 & 61 -16 ngc 7793 & 23:57:49.8 & -32:35:28 & sad & h & 230 & 3.2 & [email protected] & @xmath7718.2 & 194 . & 0.57 & 0.35 & 8.81 & & 0.3 & 14 -13 lrrcrl irac & 3.6 & 0.75 & @xmath30d@xmath44 & 1.7 & 0.02 & 4.5 & 1.02 & @xmath30d@xmath44 & 1.7 & 0.03 & 5.8 & 1.44 & @xmath30d@xmath44 & 1.7 & 0.09 & 8.0 & 2.91 & @xmath30d@xmath44 & 2.0 & 0.12 + + mips & 24 & 4.7 & @xmath30d@xmath44 & 5.7 & 0.2 & 70 & 19 . & @xmath30d@xmath44 & 17 . & 0.5 & 160 & 35 . &
@xmath30d@xmath44 & 38 . & 0.5 lcccc mips sed & 55 @xmath77 96 & 15 & 1.0 @xmath42 0.55 d@xmath44 & 1.2 mjy / sr irs sed & 14 @xmath77 40 & 62 @xmath77 124 & 0.9 @xmath42 0.55 d@xmath44 & 0.6 mjy / sr + + irs nuclear & 5.3 @xmath77 14.2 & 62 @xmath77 124 & 0.3 @xmath42 0.9 & 3 mjy & 10.0 @xmath77 37.0 & 600 & 0.3 @xmath42 0.4 & 5@xmath4210@xmath93 w / m@xmath94 + + irs extra - nuclear & 5.3 @xmath77 14.2 & 62 @xmath77 124 & 0.3 @xmath42 0.9 & 3 mjy & 10.0 @xmath77 37.0 & 600 & 0.3 @xmath42 0.4 & 5@xmath4210@xmath93 w / m@xmath94 lll irac & mosaic images & full - resolution dithered images mips image & full - map images & mips enhancer images mips sed & 1d spectra & mips enhancer spectral cubes irs & 1d spectra , line maps & 3d spectral cubes lcccrcl high - res hi maps & 21 cm & 30 & 7 & 35 & comp & vla + low - res hi maps & 21 cm & 30 & 30 @xmath77 45 & @xmath1370 & arch , sings & vla , whisp , atca + hi maps ( wsrt ) & 21 cm & 30 & 12 @xmath42 12/@xmath73 & 30 & comp & wsrt + radio continuum maps ( wsrt ) & 18 cm , 22 cm & 30 & 12 @xmath42 /@xmath73 & 30 & comp & wsrt + radio continuum maps ( vla ) & 20 cm & 30 & 30 @xmath77 45 & @xmath1360 & sings & vla + radio continuum maps ( atca ) & 20 cm & & 10 @xmath42 10/@xmath73 & 7 & sings & atca + high - res co maps & 1.3 mm & 1.7 @xmath77 3.2 & 7 & 20 & arch & bima song + low - res co maps & 1.3 mm & 3 @xmath77 5 & 50 & @xmath1340 & comp & fcrao + submillimeter maps & 850 , 450 & 2.3 @xmath77 10 & 7.5 @xmath77 14 & @xmath5920 & arch , comp & jcmt ( scuba ) + fir photometry & 170 @xmath77 240 & & 90 & 30 & arch & iso ( pht ) + @xmath65 imaging ( 2mass ) & 1.2 , 1.6 , 2.2 & @xmath95 & 2 @xmath77 3 & 75 & arch & 2mass + @xmath65 imaging ( deep ) & 1.2 , 1.6 , 2.2 & 2.5 @xmath77 8.5 & 1 @xmath77 2 & 75 & sings & steward , ctio , palomar + paschen@xmath1 , @xmath68 imaging & 1.88 , 1.6 & 0.8 & 0.2 & @xmath1350 & sings & hst ( nicmos ) + h@xmath1 imaging & 0.656 & @xmath95 & 1 @xmath77 3 & 75 & sings & kpno , ctio , steward + h@xmath1 imaging / kinematics & 0.656 & 4 @xmath77 20 & 1 @xmath77 3 & @xmath5930 & comp & cfht , eso , ohp , omm + @xmath62 imaging & 0.44 , 0.55 , 0.64 , 0.81 & @xmath95 & 1 @xmath77 2 & 75 & sings & kpno , ctio + uv imaging & 0.15 , 0.25 & @xmath95 & 3 & 27 & arch & uit + uv imaging & 0.16 , 0.24 & @xmath95 & 5 & 75 & comp & galex + x - ray imaging & 0.5 @xmath77 8 kev & 17 & 2 & @xmath5930 & arch & cxo + spectrophotometry ( strips ) & 0.36 - 0.69 & 0.9 @xmath42 0.5 @xmath15 & & 75 & sings & steward , ctio + spectrophotometry ( centers ) & 0.36 - 0.69 & 0.33 @xmath42 0.33 & & 75 & sings & steward , ctio + spectrophotometry ( nuclei ) & 0.36 - 0.69 & 0.04 @xmath42 0.04 & & 75 & sings & steward , ctio + uv spectra ( centers ) & 0.12 @xmath77 0.32 & 0.17 @xmath77 0.33 & & 25 & arch & iue + | the sirtf nearby galaxy survey is a comprehensive infrared imaging and spectroscopic survey of 75 nearby galaxies .
its primary goal is to characterize the infrared emission of galaxies and their principal infrared - emitting components , across a broad range of galaxy properties and star formation environments .
sings will provide new insights into the physical processes connecting star formation to the interstellar medium properties of galaxies , and provide a vital foundation for understanding infrared observations of the distant universe and ultraluminous and active galaxies . the galaxy sample and observing strategy
have been designed to maximize the scientific and archival value of the data set for the sirtf user community at large .
the sirtf images and spectra will be supplemented by a comprehensive multi - wavelength library of ancillary and complementary observations , including radio continuum , hi , co , submillimeter , @xmath0 , h@xmath1 , paschen-@xmath1 , ultraviolet , and x - ray data .
this paper describes the main astrophysical issues to be addressed by sings , the galaxy sample and the observing strategy , and the sirtf and other ancillary data products . |
projective configurations of points on the plane is a classical subject in algebraic geometry and its history in the context of linear systems of curves can be traced back to 18th century ( g.cramer , l.euler , etc . ) . in modern times ,
projective configurations are studied both from algebro - geometric viewpoint ( geometric invariant theory , hilbert schemes , del pezzo surfaces ) , and from combinatorial geometric viewpoint ( matroid theory ) . in the latter approach
just linear phenomena are essential , and in particular , a generic object of consideration is a _ simple @xmath4-configuration _ , that is a set of @xmath4 points in @xmath1 in which no triple of points is collinear .
the dual object is _ a simple @xmath4-arrangement _ , that is a set of @xmath4 real lines containing no concurrent triples .
a combinatorial characterization of a simple @xmath4-arrangement is its _ oriented matroid _ , which is roughly speaking a description of the mutual position of its partition polygons . for simple n - configurations
it is essentially a description how do the plane lines separate the configuration points ( see @xcite for precise definitions ) . such a combinatorial description was given for simple @xmath4-arrangements with @xmath3 in @xcite and @xcite . in the beginning of 1980s
n.mnv proved his universality theorem and in particular , constructed examples of combinatorially equivalent simple configurations which can not be connected by a deformation .
his initial example with @xmath5 was improved by p.suvorov ( 1988 ) to @xmath6 , and recently ( 2013 ) by y.tsukamoto to @xmath7 .
mnv s work motivated the first author to verify in @xcite ( see also @xcite ) that for @xmath3 the deformation classification still coincides with the combinatorial one , or in the other words , to prove connectedness of the realization spaces of the corresponding oriented matroids .
one of applications of this in low - dimensional topology was found in @xcite , via the link to the geometry of campedelli surfaces .
as @xmath4 grows , a combinatorial classification of simple @xmath4-configurations becomes a task for computer enumeration : there exist 135 combinatorial types of simple @xmath8-arrangements ( r.canham , e.halsey , 1971 , j.goodman and r.pollack , 1980 ) and 4381 types of simple @xmath9-arrangements ( j.richter-gebert , g.gonzales-springer and g.laffaille , 1989 ) .
the classification includes analysis of arrangements of _ pseudolines _ ( oriented matroids of rank @xmath10 ) , their _ stretchability _ ( realizability by lines ) and analysis of connectedness of the realization space of a matroid that gives a deformation classification ( see ( * ? ? ?
8) for more details ) .
in what follows , we need only the following summary of the deformation classification of simple @xmath4-configurations for @xmath3 .
for @xmath11 it is trivial : simple @xmath12-configurations form a single deformation component , denoted by @xmath13 .
this is because the points of such a configuration lie on a non - singular conic . for @xmath14
there are @xmath15 deformation classes shown on figure [ graphlg56 ] . on this figure , we sketched configurations @xmath16 together with some edges ( line segments ) joining pairs of points , @xmath17 .
namely , we sketch such an edge if and only if it is not crossed by any of the lines connecting pairs of the remaining @xmath18 points of @xmath16 .
the graph , @xmath19 , that we obtain for a given configuration @xmath16 will be called the _ adjacency graph _ of @xmath16 ( in the context of the oriented matroids , there is a similar notion of _ inseparability graphs _ ) . for @xmath14 , the number of its connected components
, @xmath20 , @xmath21 , @xmath10 , or @xmath22 , characterizes @xmath16 up to deformation .
the deformation classes of 6-configurations with @xmath23 components are denoted @xmath24 , @xmath25 , and the configurations of these four classes are called respectively _ cyclic , bicomponent , tricomponent _ , and _ icosahedral _ 6-configurations . given a simple 7-configuration @xmath16 , we label a point @xmath26 with an index @xmath27 if @xmath28 .
count of the labels gives a quadruple @xmath29 , where @xmath30 is the number of points @xmath26 with @xmath31 .
we call @xmath32 the _ derivative code _ of @xmath16 .
there exist 11 deformation classes of simple 7-configurations that are shown on figure [ graphlg7 ] , together with their adjacency graphs and labels @xmath33 .
+ + it is trivial to notice that if @xmath17 are adjacent vertices in graph @xmath34 , then @xmath35 , so , on figure [ graphlg7 ] we label whole components of @xmath34 rather than its vertices .
the derivative codes happen to distinguish the deformation classes , and we denote by @xmath36 the class formed by simple @xmath0-configurations @xmath16 with the derivative code @xmath37 .
problems related to the linear systems ( pencils and nets ) of real cubic curves along with related problems on the real del pezzo surfaces of degrees 1 , 2 and 3 lead to a necessity to refine the notion of simple configurations by taking into account also _
quadratic degenerations _ of configurations , in which six points become _ coconic _
( lying on one conic ) .
these problems involve also real aronhold sets of bitangents to quartics , see section [ concluding ] .
it is noteworthy that a similar motivation ( interest to aronhold sets ) was indicated by l.cummings , although in her research @xcite she did not step beyond simple 7-configurations .
the object of our interest is not that well - studied as simple configurations , although definitely is not new .
it appears for example in @xcite in the context of studying cayley octads and their relation to the aronhold sets .
we adopt here the terminology from @xcite and say that an @xmath4-configuration is _ typical _ , if it is simple and in addition does not contain coconic sextuples of points . analyzing the combinatorics of the root system @xmath38 related to the del pezzo surfaces associated with typical 7-configurations , j. sekiguchi @xcite found 14 types of such configurations ( these types give some kind of a combinatorial classification ) . later in a joint work with t.fukui ( in 1998 ) , he presented a similar computer - assisted enumeration for typical 8-configurations by analysis of the root system @xmath39 . in a different form , in terms of separation of configuration points by conics ,
a description of typical 7-configurations was given by s. le touz @xcite in the context of studying the real rational pencils of cubics .
at the same time , a similar combinatorial description of typical 7-configurations was given in @xcite .
the principal goal in @xcite ( and in the current presentation of its results ) is however to give a more subtle _ deformation classification _ of such configurations , that is to show that the realization space for each of the 14 combinatorial types of typical @xmath0-configuration is connected . like deformation classification of simple configurations
, this result leads to a deformation classification of certain associated real algebro - geometric objects . in the case of typical @xmath0-configurations
such objects are real del pezzo surfaces of degree 2 ( marked with @xmath0 exceptional curves ) , nets of cubics in @xmath40 , cayley octads and nets of quadrics in @xmath41 .
we tried to indicate one of such applications in the end of the paper . for us , a _ deformation of @xmath4-configuration _ is simply a path in the corresponding configuration space , or in the other words , a continuous family @xmath42 , @xmath43 $ ] , formed by @xmath4-configurations .
we call it _ l - deformation _ if @xmath42 are simple configurations , and _ q - deformation _ if @xmath42 are typical ones .
it is not difficult to observe ( see section 2 ) that for @xmath22-configurations the two classifications coincide : typical 6-configurations can be connected by an l - deformation if and only if they can be connected by a q - deformation . however , for @xmath44 , one l - deformation class may contain several q - deformation classes , and our main goal is to find their number in the case of @xmath45 , for each of the 11 l - deformation classes shown on figure [ graphlg7 ] . [ 14classes ] typical @xmath0-configurations split into @xmath2 q - deformation classes . among these classes , two
are contained in the class @xmath46 , three in the class @xmath47 , and each of the remaining @xmath9 l - deformation classes of simple @xmath0-configurations contains just one q - deformation class .
how to subdivide l - deformation classes @xmath46 and @xmath47 into 2 and respectively 3 q - deformation classes is shown in subsections [ non - collabsible ] and , [ decoadj ] .
in section [ preliminaries ] we recall the scheme of l - deformation classification from @xcite and give q - deformation classification of typical @xmath22-configurations .
we treat also three cases of @xmath0-configurations , in which connectedness of the realization spaces is obvious .
sections 35 are devoted to q - deformation classification for the three existing types of 7-configurations : heptagonal , hexagonal , and pentagonal . in the last section , we discuss some applications including a description of the 14 real aronhold sets ( figure [ cac7graph ] ) .
we indicated how this description can be derived from our results using cremona transformations .
we sketched also an application : a method ( alternative to that of @xcite ) to describe the topology of real rational cubics passing through the points of a 7-configuration .
this paper is essentially based on @xcite , which was partially motivated by our attempts to understand and develop the results of @xcite .
l - deformation monodromy group _ of a simple @xmath4-configuration @xmath16 we mean the subgroup , @xmath48 , of the permutation group @xmath49 realized by l - deformations , that is the image in @xmath49 of the fundamental group of the l - deformation component of @xmath16 ( using some fixed numeration of points of @xmath16 , we can and will identify @xmath49 with the symmetric group @xmath50 ) . for a typical @xmath4-configuration @xmath16 , we similarly define the _ q - deformation monodromy group _ @xmath51 formed by the permutations realized by q - deformations . in the case
@xmath52 , any permutation can be realized by a deformation ( and in fact , by a projective transformation ) , so we have @xmath53 . for @xmath11
, we obtain the dihedral group @xmath54 associated to the pentagon @xmath19 ( as it was noted the adjacency graph is an l - deformation invariant ) . more generally , we can consider a class of simple _
@xmath4-gonal @xmath4-configurations _ , @xmath16 , that are defined as ones forming a convex @xmath4-gon in the complement @xmath55 , of some line @xmath56 .
for @xmath57 this @xmath4-gon ( that coincides with the adjacency graph @xmath19 ) is preserved by the monodromy group action , and it is easy to conclude that @xmath58 . in particular , @xmath59 for @xmath60 .
it is also not difficult to show ( see @xcite ) that for @xmath22-configurations @xmath16 from components @xmath61 , @xmath62 , and @xmath63 , groups @xmath48 are respectively @xmath64 , @xmath65 , and the icosahedral group .
these facts however are not used in our paper .
the @xmath67 lines passing through the pairs of points of a simple @xmath4-configuration @xmath16 divide @xmath1 into polygons that we call _ l - polygons associated to @xmath68_. group @xmath48 acts naturally on the set of those l - polygons that can not be collapsed in a process of l - deformation .
it is easy to check that for @xmath11 none of the 31 l - polygons can be collapsed .
thus , we obtain an action of @xmath69 , which divide the set of @xmath70 l - polygons into 6 orbits : three _ internal orbits _ formed by l - polygons lying inside pentagon @xmath19 and three _ external _ ones , placed outside @xmath19 ( see figure [ 4lpolygons ] ) . by adding a point @xmath71 in one of the l - polygons associated to @xmath16 we obtain a simple @xmath22-configuration , @xmath72 .
the l - deformation class of @xmath73 depends obviously only on the @xmath48-orbit of the l - polygon containing @xmath71 .
figure [ 4lpolygons ] shows the correspondence between the six orbits and the four classes @xmath74 .
note that for @xmath23 equal to @xmath20 and @xmath22 class @xmath74 is represented by one orbit , while for @xmath21 and @xmath10 such class is represented by two orbits .
this is because @xmath48 acts transitively on the points of @xmath16 for @xmath75 , while for @xmath76 the vertices of @xmath16 split into two @xmath48-orbits . in the dual projective plane @xmath77 , consider the arrangement of lines @xmath78 which is dual to a given @xmath4-configuration @xmath79 . lines @xmath80
divide @xmath77 into polygons that we call _ subdivision polygons _ of @xmath81 .
we define the _ polygonal spectrum _ of an @xmath4-configuration @xmath16 as the @xmath82-tuple @xmath83 , where @xmath84 is the number of @xmath85-gonal subdivision polygons of @xmath81 .
euler s formula easily implies that @xmath86 for simple @xmath4-configurations .
it is easy to see also that @xmath87 if @xmath57 ( in fact , it is known that @xmath88 ) , which implies that for @xmath57 , at least one subdivision polygon has @xmath12 or more sides . if @xmath89 is a line , then point @xmath90 that is dual to @xmath91 should belong to one of the subdivision polygons , say @xmath92 . then @xmath92 is an @xmath93-gon if and only if the convex hull , @xmath94 , of @xmath16 in the affine plane @xmath95 is an @xmath93-gon .
note that @xmath93-gons @xmath92 and @xmath94 are dual : points @xmath96 are dual to lines @xmath97 disjoint from @xmath92 ( and vice versa ) .
this gives two options for a simple @xmath22-configuration @xmath16 .
the first option is @xmath98 that implies @xmath60 .
the second option is @xmath99 , @xmath100 , which means that the convex hull of @xmath16 is a pentagon in some affine chart @xmath95 .
a simple @xmath0-configurations is called _ heptagonal _ if @xmath101 , _ hexagonal _ if @xmath102 and @xmath98 , and _ pentagonal _ if @xmath103 and @xmath100 .
note that @xmath104 for any simple 7-configuration , and so , one of these three conditions is satisfied . in terms of the affine chart @xmath105 , these three cases give ( if point @xmath106 is chosen inside a subdivision polygon with the maximal number of sides ) : @xmath0 points forming a convex heptagon , @xmath22 points forming a convex hexagon plus a point inside it , and @xmath12 points forming a convex pentagon plus two points inside it ( see table below ) . [ spectrandcodes ] following @xcite , we outline here the l - deformation classification in the most essential for us case of simple @xmath0-configurations with @xmath107 . by definition ,
such configurations can be presented as @xmath72 , where @xmath16 is its cyclic 6-subconfiguration and @xmath71 is an additional point ( there are precisely @xmath108 ways to choose such a decomposition of @xmath73 ) .
like in the case of 6-configurations considered above , the monodromy group @xmath59 of @xmath60 acts on the l - polygons associated to @xmath16 ( see figure [ 8lregions ] ) , and the l - deformation class of @xmath73 depends only on the orbit ( @xmath109,[b],\dots$ ] ) of the l - polygon that contains point @xmath71 .
an additional attention is required to the four l - polygons that can be collapsed , namely , the `` central '' triangle @xmath110 in figure [ 8lregions ] and three other triangles , one of which is marked as @xmath92 on this figure : it is bounded by a principal diagonal and the two sides of the hexagon that have no common vertices with this diagonal .
we skip here the arguments from @xcite ( but after passing to the q - deformation classification in propositions [ specialqdeformation ] and [ f - connect ] , we will provide in fact a more subtle version of this proof ) .
note that @xmath111-action is well - defined on the contractible l - polygons as well as on the non - contactable ones : this action preserves @xmath110 invariant and naturally permutes the three polygons of type @xmath92 so that they form a single orbit , @xmath112 $ ] .
it is easy to see that @xmath113 for any simple 7-configuration , and that our assumption on @xmath73 admits three options .
the first option is @xmath114 , that is to say , @xmath73 is a heptagonal configuration .
this correspond to location of @xmath71 inside one of the six l - polygons from the @xmath111-orbit @xmath109 $ ] of a triangle @xmath115 .
the second option is location of @xmath71 inside hexagon @xmath19 , in one of the _ internal l - polygons _ from the orbits @xmath116 $ ] , @xmath117 $ ] , @xmath118 $ ] , or @xmath119 $ ] ( see figure [ 8lregions ] ) . in this case ,
@xmath102 and @xmath120 , that is to say , configuration @xmath73 is hexagonal . in the remaining case , @xmath71 lies outside @xmath19 , but not inside one of the six triangles of type @xmath115 .
then @xmath73 may be either hexagonal with @xmath121 , or pentagonal with @xmath122 .
totally , we enumerated 8 l - deformation classes out of 11 .
the remaining 3 pentagonal classes with @xmath123 can be described by placing two points inside a convex affine 5-configuration , as it can be understood from figure [ graphlg7 ] ( for details see @xcite ) . given a typical @xmath22-configuration @xmath16
, we say that its point @xmath26 is _ dominant _ ( _ subdominant _ ) if it lies outside of ( respectively , inside ) conic @xmath124 that passes through the remaining @xmath12 points of @xmath16 . here , by points _
inside _ ( _ outside of _ ) @xmath124 we mean points lying in the component of @xmath125 homeomorphic to a disc , ( respectively , in the other component ) .
we color the vertices of adjacency graph @xmath19 : the dominant points of @xmath16 in black and subdominant ones in white , see figure [ adjgraph6 ] for the result .
+ graphs @xmath19 are bipartite , , adjacent vertices have different colors .
[ sd ] for a typical @xmath22-configuration @xmath16 , every edge of @xmath19 connects a dominant and a subdominant points .
it follows from analysis of the pencil of conics passing through @xmath15 points of @xmath16 : a singular conic from this pencil can not intersect an edge of @xmath19 connecting the remaining two points .
let @xmath127 denote the subset formed by simple @xmath4-configurations , which have a subconfiguration of six points lying on a conic .
then @xmath128 is the set of typical @xmath4-configurations .
note that @xmath129 , so , @xmath24 for @xmath130 are formed entirely by typical configurations and thus , give three q - deformation components @xmath131 , @xmath130 , of typical @xmath22-configurations .
it follows immediately also that @xmath132 for @xmath16 from these three q - deformation components . to complete the classification it is left to observe connectedness of @xmath133 , which implies that @xmath134 is the remaining q - deformation component in @xmath135 .
connectedness follows immediately from the next lemma and the fact that any @xmath22-configuration @xmath136 has a dominant point ( in fact , it has exactly three such points , see figure [ adjgraph6 ] ) .
[ quadraticsix ] consider two hexagonal @xmath22-configurations @xmath137 , with marked dominant points @xmath138 , @xmath139 .
then , there is a @xmath126-deformation @xmath140 , @xmath43 $ ] that takes @xmath141 to @xmath142 . the same idea as in subsection [ lgc6 ]
is applied : the triangular l - polygons marked by @xmath20 on figure [ 4lpolygons ] are divided into pairs of _
q - regions _ by the conic passing through the vertices of a pentagon .
a sixth point placed outside ( inside ) of the conic is dominant ( subdominant ) .
the monodromy group @xmath143 , @xmath139 , acts transitively on the q - regions of the same kind ( in our case , on the parts of triangles marked by 1 that lie outside the conic ) , and these regions can not be contracted in the process of l - deformation of the pentagon .
therefore , an l - deformation between @xmath144 and @xmath145 that brings the q - region containing @xmath141 into the one containing @xmath142 can be extended to a required q - deformation @xmath140 ( see figure [ regionfor6pts ] ) .
it follows easily that @xmath146 , for @xmath147 , namely , @xmath65 is a subgroup of @xmath148 that preserves the colors of vertices of graph @xmath19 on figure [ adjgraph6 ] .
let @xmath149 , where @xmath37 is one of the 11 derivative codes of @xmath0-configurations ( see figure [ graphlg7 ] or table [ spectrandcodes ] ) .
like in the case of 6-configurations , some of the l - deformation components @xmath36 , namely , the ones with @xmath123 are disjoint from @xmath150 , therefore in these cases @xmath151 are q - deformation components .
from table [ spectrandcodes ] , this holds for @xmath37 being @xmath152 , @xmath153 , and @xmath154 .
we shall prove in subsection [ heptagonal - qdeformations ] connectedness of space @xmath155 formed by heptagonal typical configurations . for
a fixed configuration @xmath156 and any pair of points @xmath17 let us denote by @xmath157 the conic passing through the other five points of @xmath16 .
for each @xmath26 let us denote by @xmath158 the number of points @xmath159 for which @xmath71 lies outside of the conic @xmath157 ; this number @xmath160 will be called the _ dominance index _ of @xmath71 . the crucial fact for proving connectedness of @xmath155
is existence of a point @xmath26 such that @xmath161 .
we shall prove more : among the 14 ways to numerate cyclically the vertices of heptagon @xmath19 ( starting from any vertex , one can go around in two possible directions ) , one can distinguish a particular one that we call the _ canonical cyclic numeration_. [ cyclicorder ] for any @xmath162 , there exists a canonical cyclic numeration of its points , @xmath163 , such that @xmath164 is @xmath85 for odd @xmath85 and @xmath165 for even . in the other words ,
the sequence of @xmath166 is @xmath167 .
one can derive this proposition from the results of ( * ? ? ?
* sec.2.1 ) , but we give below a proof based on different ( in our opinion , more transparent ) approach .
the first step of our proof is the following observation .
[ unique ] for any @xmath162 , there exists at most one point @xmath26 with the dominance index @xmath161 and at most one with @xmath168 .
assume that by contrary , @xmath169 for @xmath17 .
then @xmath71 and @xmath170 are dominant points in @xmath22-configuration @xmath171 for any @xmath172 . but
dominant and subdominant points in hexagon @xmath173 are alternating ( see figure [ adjgraph6 ] ) , and so , the parity of the orders of dominant points , with respect to a cyclic numeration of the hexagon vertices , is the same . by an appropriate choice of @xmath174 ,
this parity however can be made different , which lead to a contradiction . in the case
a proof is similar .
let us fix any cyclic numeration @xmath177 of points of @xmath162 .
we denote by @xmath176 the conic passing through the points of @xmath16 different from @xmath178 and @xmath179 and put @xmath180 if @xmath178 lies inside conic @xmath176 and @xmath181 if outside , @xmath182 , @xmath183 . by definition , we have @xmath184 in what follows we apply `` modulo 7 '' index convention in notation for @xmath178 , @xmath176 and @xmath185 , that is put @xmath186 if @xmath187 , @xmath188 if @xmath189 , etc .
[ indexcalculation ] assume that @xmath190
. then 1 .
@xmath191 for all @xmath192 , @xmath193 .
2 . @xmath194 and @xmath195 provided @xmath196 .
@xmath197 provided @xmath196 .
\(a ) follows from lemma [ sd ] applied to @xmath198 .
assume that ( b ) does not hold , say , @xmath199 and @xmath200 ( the other case is analogous ) .
this means that @xmath178 lies outside of conic @xmath201 and @xmath202 lies inside . since @xmath203
, there is another conic , @xmath176 containing @xmath178 inside .
this contradicts to the bezout theorem , since @xmath201 and @xmath176 have @xmath15 common points @xmath204 , @xmath205 , @xmath206 , and in addition one more point as it is shown on figure [ examplemutual ] .
for proving ( c ) we apply lemma [ sd ] to the cyclic 6-configuration @xmath207 , in which points @xmath208 and @xmath202 become consecutive , and thus , one and only one of them is dominant , say @xmath208 ( the other case is analogous ) . then @xmath209 and @xmath200 , and thus , @xmath210 as it follows from ( b )
we say that an edge @xmath211 $ ] of heptagon @xmath19 is _ internal _
( respectively , _ external _ ) if its both endpoints lie inside ( respectively , outside of ) conic @xmath212 , or in the other words , if @xmath213 ( respectively , if @xmath214 ) .
if one endpoint lies inside and the other outside , we say that this edge is _ special _ ( see figure [ edecoration]a c ) .
[ extremal ] a special edge of the heptagon @xmath19 should connect a vertex of dominance index @xmath215 with a vertex of index @xmath22 , and in particular , such an edge is unique if exists .
the internal and external edges are consecutively alternating .
in particular , a special edge must exist ( since the number of edges is odd ) .
we sketch the internal and external edges of @xmath216 respectively as thin and thick ones .
the special edge is shown dotted and directed from the vertex of dominance index @xmath215 to the one of dominance index @xmath22 .
corollary [ extremal ] means that graph @xmath216 decorated this way should look like is shown on figure [ edecoration]d .
[ 5or7 ] the sum @xmath217 is @xmath12 if edge @xmath211 $ ] is internal , and is @xmath0 if external . by ( [ sumidx ] ) , we have @xmath218 where by lemma [ indexcalculation](b ) , @xmath219 is @xmath215 if edge @xmath211 $ ] is internal and @xmath21 if external . by lemma [ indexcalculation](a ) ,
the remaining sum is @xmath12 .
lemma [ 5or7 ] together with corollary [ extremal ] let us recover the whole sequence @xmath220 , @xmath221 , from any value different from @xmath215 and @xmath22 ( which exists by lemma [ unique ] ) , see figure [ edecoration]e .
this proof is similar to the proof of connectedness of @xmath134 in subsection [ 6confs ] .
given @xmath223 , assume that its points @xmath177 have the canonical cyclic numeration , and consider subconfiguration @xmath224 . as we observed in subsection [ simple7 ] ,
point @xmath225 lies in a triangular l - polygon @xmath115 associated with @xmath226 ( see figure [ 8lregions ] ) .
such a triangle @xmath115 is subdivided into 6 or 7 _ q - regions _ @xmath227 by the conics @xmath228 , @xmath229 , that connect quintuples of points of @xmath226 , see figure [ adjgraph6 ] .
the only one of these q - regions that can be collapsed is @xmath230 , so , the monodromy group @xmath231 acts on the q - regions of types @xmath227 , @xmath232 and clearly , forms 6 orbits , denoted respectively by @xmath233 $ ] .
our choice of @xmath225 with @xmath234 means that this point lies inside region @xmath235 .
transitivity of @xmath65-action on the q - regions of type @xmath235 and impossibility for such a region to be collapsed in the process of a q - deformation implies connectedness of @xmath222 . placing point @xmath225 in a q - region @xmath227 , @xmath236 , instead of @xmath235 lead to a new canonical cyclic ordering of the points of @xmath16 ( different from @xmath177 ) . to recover that order , it is sufficient to know the two points , of dominance indices @xmath215 and @xmath22 .
table [ table ] shows how this pair of points depend on the region @xmath227 .
in all the cases we follow the same scheme of q - deformation classification as for heptagonal configurations in subsection [ heptagonal - qdeformations ] . namely , we consider a typical 7-configuration with a marked point @xmath237 , so that @xmath238 ( a cyclic 6-subconfiguration ) . in this section
we assume that @xmath225 lies inside hexagon @xmath239 , which corresponds to the case of hexagonal configuration @xmath16 ( see subsection [ dual ] ) . for a given @xmath16
the number of such choices of @xmath225 is equal to @xmath240 ( recall that an affine chart @xmath241 in which the convex hull of @xmath16 is hexagonal corresponds in the dual terms to a choice of point @xmath106 inside a hexagonal component of @xmath242 for the dual arrangement @xmath81 ) . since in our case @xmath243
we conclude that a choice of marked point @xmath225 is unique . in the next subsection
we consider @xmath225 lying outside @xmath239 in one of l - polygons that correspond to pentagonal configurations ( so , we exclude previously considered cases of heptagonal and hexagonal 7-configurations ) .
conics @xmath244 passing through the points of 5-subconfigurations @xmath245 , @xmath229 , can subdivide an l - polygon into several _
q - regions _ like in subsection [ heptagonal - qdeformations ] , and our aim is to analyze which of these regions can not be contracted in a process of q - deformation , and how the monodromy group @xmath246 does act on them . we always choose a cyclic order of points @xmath247 so that @xmath248 is dominant ( then @xmath249 and @xmath250 are dominant too , whereas @xmath251 , @xmath252 , @xmath253 are subdominant ) . then conics @xmath254 , @xmath255 , and @xmath256 contain hexagon @xmath239 inside , whereas @xmath257 , @xmath258 , and @xmath259 intersect the internal l - polygons of @xmath226 , see figure [ qregions ] . on figure [ qregions ] conics @xmath257 , @xmath258 , @xmath259 do not intersect the sides of the hexagon . if they do intersect , then the shape of q - regions of types @xmath260 and @xmath261 may change , which changes figure [ qregions ] a bit , but not essentially for our arguments . the internal l - polygons of types @xmath262 and @xmath110 are obviously contained inside these conics and only l - polygons of types @xmath263 and @xmath264 are actually subdivided into q - regions .
namely , the latter l - polygons are subdivided into q - regions @xmath260 , @xmath261 and respectively @xmath265 , @xmath266 , @xmath267 as it is shown .
next , we can easily see that monodromy group @xmath231 acts transitively on the q - regions of each type , which gives seven @xmath65-orbits : @xmath260 , @xmath261 , @xmath265 , @xmath266 , @xmath267 , @xmath262 and @xmath110 .
note that triangle @xmath110 is the only internal q - region of @xmath268 that can be collapsed by a q - deformation .
thus , any pair of hexagonal @xmath0-configurations @xmath73 and @xmath269 whose marked points , @xmath270 and @xmath271 belong to the same @xmath65-orbit of the internal q - regions different from @xmath110 can be connected by a q - deformation .
namely , we start with a q - deformation between @xmath272 and @xmath273 that transforms the q - region containing @xmath270 to the one containing @xmath271 and extend this deformation to the seventh points using non - contractibility of the given type of q - regions .
this yields q - deformation classes @xmath274 , @xmath275 , @xmath276 , @xmath277 , @xmath278 , and @xmath279 that correspond respectively to the @xmath65-orbits of types @xmath260 , @xmath261 , @xmath265 , @xmath266 , @xmath267 , and @xmath262 , see figure [ qregions ] .
consider a subset @xmath280 formed by typical hexagonal 6-configurations @xmath281 whose _ principal diagonals _
@xmath282 , @xmath283 , and @xmath284 are not concurrent , or in the other words , whose l - polygon @xmath110 is not collapsed .
connectedness of space @xmath285 formed by 7-configurations @xmath237 with @xmath225 placed in the q - region @xmath110 would follow from connectedness of @xmath286 . [ specialqdeformation ] space @xmath286 is connected .
given a pair of configurations , @xmath287 , @xmath139 , we need to connect them by some deformation @xmath288 , @xmath43 $ ] .
let us choose a cyclic numeration of points , @xmath289 , @xmath139 , so that @xmath290 are dominant .
at the first step , we can achieve that the triangular q - regions `` e '' of the both configurations coincide , so that the dominant and subdominant points in @xmath291 go in the same order , as it is shown on figure [ centraltrianglesame](a ) , that is , points @xmath292 , @xmath139 , lie on the ray that is the extention of side @xmath293 ( then the other rays extending the sides if triangle @xmath294 are also of the same color ) .
this can be done by a projective transformation sending the diagonals @xmath295 , @xmath296 , @xmath297 , and the _ infinity line _ , @xmath298 , ( that pass in the complement of @xmath299 ) to the corresponding diagonals and the `` infinity line '' for configuration @xmath300 ( existence of a deformation is due to connectedness of @xmath301 ) .
if the mutual positions of the dominant and subdominant points on the lines in @xmath302 and @xmath300 will differ , then it can be made like on figure [ centraltrianglesame](a ) by a projective transformation that permutes the three diagonals while preserving @xmath298 .
( one can also use flexibility of the initial numeration of vertices in @xmath300 ) . fixing a triangle @xmath303 ,
let us denote by @xmath304 the subspace of @xmath305 consisting of hexagonal @xmath22-configurations whose dominant points lie on the affine rays that are continuations of sides @xmath293 , @xmath306 , and @xmath307 , and subdominant points lie on the continuations of @xmath308 , @xmath309 , and @xmath310 , as it is shown on figure [ centraltrianglesame](a ) .
the final step of the proof is connectedness of @xmath304 .
[ connect ] for a fixed triangle @xmath311 , the configuration space @xmath304 is connected .
consider @xmath312 , @xmath281 , where point @xmath248 is dominant one lying on @xmath293 .
consider conic @xmath126 passing trough @xmath313 .
triangle @xmath314 lies inside @xmath126 and point @xmath248 lies outside .
this gives a one - to - one correspondence between @xmath304 and the space of pairs @xmath315 , where @xmath126 is an ellipse containing @xmath294 inside and @xmath71 is a point on the continuation of @xmath293 lying outside @xmath126 ( see figure [ centraltrianglesame](b ) ) .
the space of such ellipses is connected ( and in fact , contractible ) , and the projection of @xmath304 to this space is fibration with a contractible fiber .
thus , @xmath304 is connected ( and in fact , is contractible ) .
-3 mm figure [ qhexed7 ] shows the adjacency graphs of typical hexagonal configurations @xmath16 endowed + + additionally with the _ vertex coloring _ for @xmath316 as in subsection [ qgc ] ( black for dominant and white for subdominant points ) and with the _ edge decoration _ for a connected component of @xmath34 labeled by @xmath20 .
namely , such an edge @xmath317 $ ] is _ thin _
if @xmath178 and @xmath179 lie inside conic @xmath318 , _ thick _ if they lie outside , or _
dotted and directed _ from @xmath178 to @xmath179 if @xmath178 lies inside and @xmath179 lies outside ( like is shown on figure [ edecoration ] ) .
such decoration lets us distinguish q - deformation types of hexagonal configurations .
recall that a pentagonal typical 7-configuration @xmath16 has precisely @xmath108 vertices @xmath26 such that @xmath319 .
moreover , for pentagonal configurations @xmath108 is either @xmath215 or @xmath20 ( see table 1 ) .
so , in the case @xmath107 ( that is @xmath320 ) considered in the next section such a vertex @xmath71 is unique , and we can ( and will ) color the six vertices of @xmath226 according to their dominancy as before .
a pentagonal typical @xmath0-configuration with @xmath321 ( and thus , @xmath323 ) can be presented like in the previous section as @xmath237 , where @xmath324 .
the difference is that now point @xmath225 lies outside hexagon @xmath239 .
more precisely , @xmath225 should lie in an l - polygon of type @xmath92 , or @xmath322 , or @xmath94 , since the other types of l - polygons correspond either to heptagonal ( the case of l - polygons of type @xmath115 ) or hexagonal ( the case of types @xmath325 and @xmath326 ) configurations @xmath16 that were analyzed before .
the first crucial observation is that none of the conics @xmath244 , @xmath229 can intersect these three types of external polygons , and therefore , such l - polygons are not subdivided into q - regions like in the case of types @xmath115 , @xmath263 and @xmath264 .
it is clear from figure [ pentagonregions ] : conics @xmath244 should lie in the shaded part that is formed by l - polygons of types @xmath115 , @xmath325 and @xmath326 .
the second observation is that l - polygons of each type , @xmath92 , @xmath322 , or @xmath94 , form a single orbit with respect to the action of monodromy group @xmath327 .
the third evident observation is that l - polygons of types @xmath322 and @xmath94 can not be contracted by a q - deformation ( as they can not be contracted even by an l - deformation ) .
together these observations imply that the corresponding to l - polygon types @xmath322 and @xmath94 ( see table [ spectrandcodes ] ) configuration spaces @xmath328 and @xmath329 are connected , and thus , are q - deformation components .
[ f - connectedness ] the configuration space @xmath330 is connected , or equivalently , l - deformation component @xmath331 that correspond to l - polygon of type @xmath92 contains a unique q - deformation component .
a configuration @xmath332 has a unique distinguished point @xmath225 , such that @xmath237 , where @xmath238 and @xmath225 lies in the l - polygon of type @xmath92 .
such polygon is a triangle whose vertices we denote by @xmath333 , @xmath334 , and @xmath335 using the following rule . by definition of @xmath92-type polygon ,
one of its supporting lines should be a principal diagonal passing through two opposite vertices of hexagon @xmath239 .
we can choose a cyclic numeration of points @xmath336 so that these opposite vertices are @xmath248 and @xmath252 , and @xmath248 is a dominant point ( then @xmath249 , @xmath250 are also dominant , and @xmath251 , @xmath252 , @xmath253 are subdominant ) .
two vertices of the triangle on the line @xmath282 are denoted by @xmath333 and @xmath334 in such an order that @xmath333 , @xmath334 , @xmath248 , @xmath252 go consecutively on this line , like it is shown on figure [ xyz](a ) , and the third point of the triangle is denoted by @xmath335 .
the direction of cyclic numeration of points @xmath178 can be also chosen so that points @xmath251 , @xmath249 lie on the line @xmath310 and @xmath250 , @xmath253 on @xmath306 ( see figure [ xyz](a ) ) . by a projective transformation
we can map a triangle @xmath314 to any other triangle on @xmath1 , so , in what follows we suppose that triangle @xmath314 is fixed and denote by @xmath337 the subspace formed by typical cyclic configurations having @xmath314 as its l - polygon of type @xmath92 and having a cyclic numeration of points @xmath338 satisfying the above convention . then , proposition [ f - connectedness ] follows from connectedness of @xmath339 .
[ f - connect ] for a fixed triangle @xmath314 , the configuration space @xmath339 is connected . using the same idea as in lemma
[ connect ] , we associate with a configuration @xmath340 a pair @xmath315 , where @xmath71 is the dominant point of @xmath16 on the line @xmath293 ( that is @xmath248 in the notation used above ) and @xmath126 is the conic passing through the other points of @xmath16 .
note that @xmath16 can be recovered from pair @xmath315 associated to it in a unique way .
position of @xmath126 can be characterized by the conditions that triangle @xmath314 lie outside @xmath126 and the lines @xmath293 , @xmath310 , @xmath306 intersect conic @xmath126 at two points , so that the chord of conic @xmath126 that is cut by line @xmath293 lies between the two other chords that are cut by @xmath310 and @xmath306 ( see figure [ xyz](b ) ) .
the set of conics satisfying these requirements is obviously connected ( and in fact , is contractible ) . for each conic @xmath126 like this , there is some interval on the line @xmath293 ( see figure [ xyz ] ) formed by points @xmath71 such that @xmath315 is associated to some @xmath340 .
thus , the set of such pairs @xmath315 , or equivalently @xmath339 , is also connected .
-3 mm we have shown in subsection [ 3trivial ] connectedness of three components @xmath341 , @xmath342 , and @xmath343 of pentagonal 7-configurations with @xmath123 . in subsection
[ heptagonal - qdeformations ] we have shown connectedness of the component @xmath155 formed by heptagonal typical 7-configurations , and in section [ hex ] found seven connected components formed by hexagonal 7-configurations .
the remaining 3 cases of pentagonal configurations with @xmath320 were analyzed in subsections [ g - h ] and [ f - polygons ] .
by blowing up @xmath344 at the points of a typical 6-configuration @xmath345 we obtain a del pezzo surface @xmath346 of degree 3 that can be realized by anti - canonical embedding as a cubic surface in @xmath347 .
the exceptional curves of blowing up form a configuration of six skew lines @xmath348 that is nothing but a half of schlfli s double six of lines , and we call below such @xmath349 the _ skew six of lines _ represented by @xmath16 . in the real setting , for @xmath350 , cubic surface
@xmath346 is real and _ maximal _ , where the latter means by definition that the real locus @xmath351 is homeomorphic to @xmath352 .
the four deformation classes of typical 6-configurations give four types of real skew sixes of lines : _ cyclic , bicomponent , tricomponent and icosahedral_. it was observed in @xcite that the complementary real skew six of lines ( that forms together with @xmath349 a real double six on @xmath346 ) has the same type as a given one , and so , we can speak of the _ four types of real double sixes of lines_. it was shown by v.mazurovski ( see @xcite ) that there exist 11 _ coarse deformation classes _ of six skew line configurations in @xmath353 : here _ coarse _ means that deformation equivalence is combined with projective ( possibly orientation - reversing ) equivalence , for details see @xcite . among these 11 classes ,
9 can be realized by so called _ join configurations _ , @xmath354 , that can be presented by permutations @xmath355 as follows .
fixing consecutive points @xmath356 and @xmath357 on a pair of auxiliary skew lines , @xmath358 and @xmath359 respectively , we let @xmath360 , where line @xmath361 joins @xmath178 with @xmath362 , @xmath229 .
we denote such a configuration ( and sometimes its coarse deformation class ) by @xmath363 .
the remaining two coarse deformation classes among 11 can not be represented by join configurations @xmath363 ; these two classes are denoted in @xcite by @xmath364 and @xmath365 . as it is shown in @xcite , the cyclic , bicomponent , and tricomponent coarse deformation classes of real skew sixes @xmath366
are realized as @xmath363 , where @xmath367 is respectively @xmath368 , @xmath369 , and @xmath370 , where @xmath367 is recorded as @xmath371 ( see figure [ cac6graph ] ) .
the icosahedral coarse deformation class corresponds to the class @xmath365 from @xcite . a change of cyclic orderings of points @xmath178
, @xmath372 on lines @xmath358 and @xmath359 clearly does not change the coarse deformation class of @xmath363 . in the other words ,
the coarse deformation class of @xmath363 is an invariant of the orbit @xmath373\in s_6/(\dd_6\times\dd_6)$ ] of @xmath355 with respect to the left - and - right multiplication action of @xmath374 in @xmath375 for the dihedral subgroup @xmath376 . with a permutation
@xmath377 we associate a diagram @xmath378 obtained by connecting cyclically ordered vertices @xmath379 of a regular @xmath4-gon by diagonals @xmath380 , @xmath381 , ( here , @xmath382 ) .
then `` the shape of @xmath378 '' characterizes class @xmath373\in s_n/(\dd_n\times\dd_n)$ ] , see figure [ cac6graph ] for the _ hexagrams _ representing the cyclic , bicomponent , and tricomponent permutation orbits @xmath373 $ ] , namely , @xmath383 $ ] , @xmath384 $ ] and @xmath385 $ ] . by dropping a line from a real skew six @xmath386
we obtain _ a real skew five _ , @xmath387 , that can be realized similarly , as a join configuration @xmath363 for @xmath388 .
it was shown in @xcite that the class @xmath373\in s_5/(\dd_5\times\dd_5)$ ] does not depend on the line in @xmath386 that we dropped , including the case of icosahedral real double sixes , see the corresponding _ pentagrams _
@xmath389}$ ] on figure [ cac6graph ] . by blowing up the points of a typical 7-configuration , @xmath350
, we obtain a non - singular real del pezzo surface @xmath346 of degree @xmath21 with a configuration @xmath390 of @xmath0 disjoint real lines ( the exceptional curves of blowing up ) .
the anti - canonical linear system maps @xmath346 to a projective plane as a double covering branched along a non - singular real quartic , whose real locus has 4 connected components .
each of the 7 lines of @xmath366 is projected to a real bitangent to this quartic , and the corresponding arrangement of 7 bitangents is called an _ aronhold set_. the 14 q - deformation classes of typical 7-configurations yield 14 types of real aronhold sets , which were described in @xcite , see appendix . among various known criteria to recognize that real bitangents @xmath361 , @xmath391 , to a real quartic form an aronhold set , topologically the most practical one is perhaps possibility to color the two line segments between the tangency points on each @xmath361 in two colors , so that at the intersection points @xmath392 , the corresponding line segments of @xmath361 and @xmath393 are colored differently .
such colorings are indicated on the figures in the appendix . in @xcite ,
fiedler - le - touz analyzed real nodal cubics , @xmath394 , passing through the points @xmath395 of a heptagonal configuration , @xmath396 , and having a node at one of the points @xmath397 , and described in which order the points of @xmath16 may follow on the real locus of @xmath394 ( see figure [ nodalcubic ] ) .
recently , a similar analysis was done for the other types of 7-configurations , see @xcite .
we proposed an alternative approach based on the real aronhold set , @xmath398 , corresponding to a given typical 7-configuration @xmath16 .
namely , the order in which cubic @xmath394 passes through the points @xmath179 is the order in which bitangent @xmath361 intersects other bitangents @xmath393 .
the two branches of @xmath394 at the node correspond to the two tangency points of @xmath361 .
possibility of two shapes of cubic @xmath399 shown on figure [ nodalcubic ] correspond to possibility to deform a real quartic with 4 ovals , so that bitangent @xmath400 moves away from an oval , as it is shown in the appendix on the top figure : the two tangency points to @xmath400 on that oval are deformed into two imaginary ( complex conjugate ) tangency points .
similarly , one can shift double bitangents to the same ovals in the other of real aronhold sets shown in the appendix .
the two loops ( finite and infinite ) of a real nodal cubic @xmath394 that correspond to the two line segments on @xmath361 bounded by the tangency points can be distinguished by the following parity rule .
line @xmath361 contains six points of intersection with @xmath393 , @xmath401 , @xmath402 , and one more intersection point , with a line @xmath403 obtained by shifting @xmath361 away from the real locus of the quartic .
one of the two line segments contains even number of intersection points , and it corresponds to the `` finite '' loop of @xmath394 , and the other line segment represents the `` infinite '' loop of @xmath394 . an elementary real cremona transformation , @xmath404 , based at a triple of points @xmath405 transforms a typical 7-configuration @xmath163 to another typical 7-configuration @xmath406 .
starting with a configuration @xmath156 , we can realize the other 13 q - deformation classes of 7-configurations as @xmath407 for a suitable choice of @xmath408 , as it is shown on figure [ cremona ] , see @xcite for more details .
this construction is used to produce the real aronhold sets shown in the appendix .
the 14 figures below show real aronhold sets representing typical planar @xmath0-configurations . in the case of @xmath155 on the top figure
we have shown a possible variation of one of the bitangents that has two contacts to the same oval : it can be shifted from this oval after a deformation of the quartic , so that the contact points become imaginary .
similar variations are possible in the other 9 cases ( except @xmath275 , @xmath277 , @xmath278 , and @xmath279 ) .
bjrner a. , las vergnas m. , sturmfels b. , white n. and ziegler g. _ oriented matroids_. encyclopedia of math and its appl .
46 , ( cambridge univ .
press , cambridge 1999 ) cummings l.d .
_ hexagonal system of seven lines in a plane .
_ , bull.am.math.soc . * 38 * , no 1 , 105 - 109 ( 1932 ) drobotukhina ju.v .
, o.viro _ configurations of skew lines_. algebra i analiz 1:4 , 222 - 246 ( 1989 ) , english translation in leningrad math .
j. * 1:4 * , 1027 - 1050 , upgraded in arxiv : math/0611374 ( 1990 ) fiedler - le touz s. _ pencils of cubics as tools to solve an interpolation problem_. applicable algebra in engineering , communication and computing , * 18 * , 53 - 70 ( 2007 ) .
doi:10.1007/s00200 - 006 - 0028 - 3 fiedler - le touz s. _ pencils of cubics with eight base points lying in convex position in @xmath1_. arxiv:1012.2679 [ math.ag ] fiedler - le touz s. _ rational pencils of cubics and configurations of six or seven points in @xmath1_. arxiv : 1210.7146 math.ag . finashin s. _ projective configurations and real algebraic curves _ , ph.d .
thesis , leningrad state university , pp.123 .
leningrad , russia ( 1985 ) .
( section on rigid isotopies of configurations is available in viniti ( moscow , russia ) , report 506785 ( in russian ) , 16/07/1985 , 45 pp . ) finashin s. _ configurations of seven points in @xmath353 _ , in : topology and geometry , rokhlin seminar ( ed .
o.viro ) , lect . notes math .
, * 1346 * , 501 - 526 ( 1988 ) .
doi:10.1007/bfb0082791 gross b.h . ,
harris j. _ on some geometric constructions related to theta characteristics _
contributions to automorphic forms , geometry and number theory , john hopkins press ( 2004 ) 279311 | a configuration of @xmath0 points in @xmath1 is called _ typical _ if it has no collinear triples and no coconic sextuples of points .
we show that there exist @xmath2 deformation classes of such configurations .
this yields classification of real aronhold sets . |
net - charge fluctuations are measured in nuclear collisions by many rhic and sps experiments @xcite .
conserved quantities such as net electric charge , baryon number , and strangeness can fluctuate when measured in limited rapidity intervals .
these fluctuations occur mainly because the number of produced particles varies with each collision event due to differences in impact parameter , energy deposition , and baryon stopping .
a variety of interesting dynamic effects can also contribute to these fluctuations @xcite .
in particular , fluctuations of mean @xmath0 , net charge , and baryon number may probe the hadronization mechanism of the quark - gluon plasma @xcite .
fluctuations of conserved quantities are perhaps the best probes of hadronization , because conservation laws limit the dissipation they suffer after hadronization has occurred @xcite .
this dissipation occurs by diffusion . while the effect of diffusion on charge and baryon fluctuations has been studied in refs .
@xcite , the classic diffusion equations used are problematic in a relativistic context , because they allow signals to propagate with infinite speed , violating causality .
in this paper we study the dissipation of net charge fluctuations in rhic collisions using a causal diffusion equation .
we find that causality inhibits dissipation , so that extraordinary fluctuations , if present , may survive to be detected . in sec .
[ sec : diffusion ] we discuss how the classic approach to diffusion violates causality , and present a causal equation that resolves this problem .
the derivation we include facilitates our work in later sections .
we generalize this equation for relativistic fluids in nuclear collisions in sec .
[ sec : ions ] .
related causal fluid equations for viscosity and heat conduction have been introduced in @xcite and @xcite with different heavy - ion applications in mind .
our causal formulation can be crucial for the description of net - charge fluctuations , which involve rapid changes in the inhomogeneous collision environment @xcite .
we turn to this problem in secs .
[ sec : fluct][sec : discussion ] . in sec .
[ sec : fluct ] we discuss fluctuation measurements and their relation to two - particle correlation functions . in sec .
[ sec : fluctdiff ] we introduce techniques from @xcite to compute the effect of causal and classic diffusion on these correlation functions .
we then estimate the impact of causal diffusion on fluctuation measurements in sec .
[ sec : discussion ] .
these three sections are close in spirit to work by shuryak and stephanov , where a classic diffusion model is used @xcite .
to understand why the propagation speed in classic diffusion is essentially infinite , recall that the diffusion of particles through a medium is equivalent to a random walk in the continuum limit .
the variance of the particle s displacement increases by @xmath1 in the time interval @xmath2 between random steps .
the average propagation speed @xmath3 diverges in the continuum limit , where @xmath4 with the diffusion coefficient @xmath5 held fixed .
correspondingly , a delta function density spike is instantaneously spread by diffusion into a gaussian , with tails that extend to infinity . . , width=259 ] -0.2 in a causal alternative to the diffusion equation is the telegraph equation , @xmath6 where @xmath7 is the diffusion coefficient and @xmath8 is the relaxation time for diffusion @xcite .
signals propagate at the finite speed @xmath9 , so that a delta function spike spreads behind a front travelling at @xmath10 , as shown in fig .
[ fig : causal ] .
the classic diffusion equation omits the term @xmath11 .
classic diffusion describes the matter well behind the front at times @xmath12 .
this contrasting behavior is generic of hyperbolic equations such as ( [ eq : grad ] ) , which include a second order time derivative , compared to parabolic equations , like the classic first - order diffusion equation @xcite .
causality concerns are not restricted to relativistic quarks or pions diffusing through a quark gluon plasma or hadron gas .
they are common and constantly debated whenever time varying diffusion and heat conduction phenomena are discussed @xcite . however , for most nonrelativistic systems , causality violations are minuscule , so that classic diffusion can be used .
this need not be the case for relativistic fluids produced in nuclear collisions @xcite . to motivate ( [ eq : grad ] ) and understand its limitations
, we first derive the diffusion coefficient using the boltzmann equation in the relaxation time approximation ; see , e.g. , @xcite .
equation ( [ eq : grad ] ) can also be obtained by the moment method @xcite or by sum - rule arguments @xcite . for simplicity , we focus on electric charge transport by a single species .
generalization to multiple species and other currents such as baryon number is straightforward , except for the challenging case of color transport @xcite .
we describe the evolution of the phase space distribution @xmath13 using + _
f= -(f - f_e),[eq : boltz]where @xmath14 is the relaxation time , @xmath15 , and @xmath16 . the local equilibrium distribution satisfies @xmath17\pm 1\}^{-1}$ ] where @xmath18 is the temperature @xmath18 , @xmath19 the chemical potential , and the @xmath20 or @xmath21 sign describes bosons or fermions .
suppose that @xmath19 differs from a local equilibrium value by a small sinusoidally varying perturbation @xmath22 , with @xmath23 .
then @xmath13 is driven from local equilibrium by an amount @xmath24 .
equation ( [ eq : boltz ] ) implies [ eq : domegak0 ] f ( , ) = -i _ f_e ( , ) , plus corrections of order @xmath25 .
the net charge current is ( , ) f ( , ) _
d= -i d ( , ) n ( , ) , where @xmath26 .
the diffusion coefficient is @xmath27 which is a relativistic generalization of a completely standard kinetic theory result . for @xmath28
, we recover the familiar static diffusion coefficient @xmath29 , where the thermal velocity @xmath30 for massless particles . to obtain quantitative results from this relaxation time approximation
, we identify @xmath14 with the relaxation time for diffusion @xmath8 obtained by more sophisticated methods , see e.g. @xcite . to obtain ( [ eq : grad ] ) , we omit the @xmath31 dependence in ( [ eq : domegak ] ) , to find d ( , 0)=d/(1- i_d ) , [ eq : diffomega]where @xmath7 is the static diffusion coefficient and @xmath32 .
we then write ( 1- i _ d)(,)= - i d n(,),[2]to find [ eq : cattaneo ] _ d ( , t ) + ( , t ) = - d n(,t ) , [ 3 ] the maxwell - cattaneo relation @xcite . combining ( [ 3 ] ) with current conservation @xmath33 yields ( [ eq : grad ] ) .
we will extend this argument in sec .
[ sec : fluctdiff ] to derive eq .
( [ eq : vardiffcaus ] ) .
classic diffusion follows from ( [ 3 ] ) when the @xmath8 term is negligible , a result known as fick s law .
fick s law violates causality because any density change instantaneously causes current to flow .
including the @xmath8 term is the simplest way to incorporate a causal time lag for this current response @xcite .
moreover , ( [ 3 ] ) is self consistent in that it saturates the @xmath13-sum rule @xcite .
however , while ( [ eq : grad ] ) is a plausible approximation , the @xmath34 limit ( [ 2 ] ) is not strictly justified .
possible generalizations can include the nonlocal equations or gradient expansions derived from ( [ eq : domegak0 ] ) and ( [ eq : domegak ] ) .
if ( [ eq : grad ] ) leads to substantial corrections to classic diffusion , then such generalizations can be worth considering . in solving ( [ eq : grad ] )
we must impose initial conditions on the current that respect causality .
suppose that we introduce a density pulse at @xmath35 at time @xmath36 .
microscopically , particles begin to stream freely away from this point .
the current implied by ( [ eq : cattaneo ] ) is @xmath37 scattering with the surrounding medium eventually establishes a steady state in which fick s law holds , as we see from ( [ eq : cattaneo ] ) for @xmath38 , but this takes a time @xmath39 .
an assumption of no initial flow @xmath40 is consistent with our physical picture , since there is no preferred direction for the initial velocity of each particle .
note that an alternative choice @xmath41 would imply that the current always follows fick s law ; the corresponding solutions of ( [ eq : grad ] ) would never differ appreciably from classic diffusion .
however , this choice is not causal because it requires that particles `` know '' about the medium before they have had any opportunity to interact with it .
we now extend ( [ eq : grad ] ) to study the diffusion of charge through the relativistic fluid produced in a nuclear collision .
the fluid flows with four velocity @xmath42 determined by solving the hydrodynamic equations @xmath43 together with the appropriate equation of state .
we choose the landau - lifshitz definition of @xmath42 in terms of momentum current @xcite .
for sufficiently high energy collisions , the relative concentration of net charge is small enough that it has no appreciable impact on @xmath44 .
correspondingly , we take @xmath44 to be a fixed function of @xmath45 and @xmath46 . following @xcite , we define the co - moving time derivative and gradient d_u^_and ^=^- u^u^ _ [ eq : cograd]for the metric @xmath47 . in the local rest frame where @xmath48 , these quantities are the time derivative and @xmath49 , where @xmath50 is the three - gradient .
the total charge current in the moving fluid is @xmath51 , where the first contribution is due to flow and the second to diffusion .
continuity then implies [ eq : continuity ] _
( n u^)=- _
j^. we assume that the diffusion current satisfies _
d d_j^+j^= d ^ n [ eq : maxwellcattaneo],which reduces to ( [ 3 ] ) in the local rest frame . to illustrate the effect of flow on diffusion , we consider longitudinal bjorken flow , @xmath52 , where @xmath53 , @xmath54 , the density is a function only of @xmath55 and the current is @xmath56 @xcite .
the continuity equation ( [ eq : continuity ] ) is then ( + ) n = ( n ) = - _ j^.[eq : contf]to evaluate the covariant maxwell - cattaneo relation , we differentiate ( [ eq : maxwellcattaneo ] ) to find _ d _ ( u^ _
j^ ) + _ j^ = - d ^2 n , [ 20]where @xmath57 .
we then write @xmath58 where the second line follows from eqs .
( 17 ) and ( 20 ) of ref .
then ( [ 20 ] ) and ( [ 22 ] ) imply @xmath59 together , ( [ eq : contf ] ) and ( [ eq : mcf ] ) describe causal diffusion . to obtain an equation analogous to ( [ eq : grad ] ) for the expanding system ,
observe that the rapidity density @xmath60 , where @xmath61 is the transverse area of the two colliding nuclei . if one identifies spatial rapidity @xmath62 with the momentum - space rapidity of particles
, then @xmath63 is observable .
we combine ( [ eq : contf ] ) and ( [ eq : mcf ] ) to find that this rapidity density satisfies _
d + = [ 25],where @xmath64 for longitudinal expansion .
initial conditions for @xmath63 and @xmath65 at the formation time @xmath66 must be specified . in view of the causality argument surrounding ( [ eq : catint ] ) ,
we assume that the initial diffusion current @xmath67 , so that ( [ eq : contf ] ) implies @xmath68 at @xmath69 . note that the total current @xmath70 is initially non - zero , since the underlying medium is not at rest . in the absence of diffusion ,
longitudinal expansion leaves @xmath63 fixed , as we see from ( [ eq : contf ] ) for @xmath71 .
diffusion tends to broaden the rapidity distribution . to characterize this broadening , we compute the rapidity width defined by @xmath72 , where @xmath73 and @xmath74 . we multiply both sides of ( [ 25 ] ) by @xmath75 and integrate to find _
[ 27]observe that classic diffusion follows from ( [ 27 ] ) for @xmath76 with @xmath7 fixed . in that case
the width increases by [ 30 ] v= ( 1- ) , where @xmath77 .
the rapidity width indeed increases , but the competition between longitudinal expansion and diffusion limits this increase to an asymptotic value @xmath78 .
we now solve ( [ 27 ] ) for causal diffusion to obtain [ 28 ] v = _
1^/_o f ( , ) d , where @xmath79 and f(,)=_1^^-2e ^(- ) d.[29]we have taken @xmath80 at @xmath81 , as required when the initial @xmath65 vanishes .
figure [ fig:1 ] compares @xmath82 for classic and causal diffusion .
we find that the rapidity broadening is always slower for causal diffusion compared to the classic case .
these solutions approach one another for @xmath83 ; they are within 20% for @xmath84 for @xmath85 . the classic result ( [ 30 ] )
follows from ( [ 28 ] ) and ( [ 29 ] ) for @xmath86 . ) and ( [ 30 ] ) respectively .
causal curves are for @xmath87 , 1 , and 1.5 .
rapidity spreads are divided by the asymptotic value @xmath88.,width=288 ] we have so far assumed that @xmath8 and @xmath7 are constant , but this may not be the case . these coefficients can vary with the overall density as the system expands and rarefies . as an alternative extreme ,
suppose that diffusion occurs as massless charged particles elastically scatter with the expanding fluid of density @xmath89 .
kinetic theory implies that @xmath90 and @xmath91 .
if we take the scattering rate @xmath92 averaged over species and temperature to be constant , but assume @xmath93 as determined by entropy conservation , then @xmath94 .
realistically , particle - mass effects would reduce the rate of growth of @xmath8 and @xmath7 , as would the temperature dependence of @xmath95 .
nevertheless , the linear growth rate is worth considering to illustrate how rarefaction can change the results . including this extreme effect of rarefaction implies @xmath96 , so
that _ d = /,[eq : taudt]where we now fix the parameter @xmath97 at the initial time . taking @xmath98 gives @xmath99 so that the diffusion equation ( [ 27 ] ) becomes @xmath100 as before , classic diffusion is obtained by omitting the second derivative term .
we find [ eq : classict]v = .
we see that the width now increases without bound , which is not surprising since our time varying parameters ( [ eq : taudt ] , [ eq : dt ] ) imply that @xmath101
. informed by the classic case , we write the causal equation as @xmath102 where @xmath103 .
when @xmath104 , the first derivative vanishes , so that @xmath105 for @xmath106 , we find @xmath107\right\}.\ ] ] for @xmath108 , the increase of @xmath109 is always smaller than classic diffusion ( [ eq : classict ] ) .
in this regime , we see that @xmath110 as @xmath111 .
the difference of the ratio from unity in the long time limit reflects the fact that the rarefaction rate @xmath112 and diffusion rate @xmath113 are in fixed proportion for all time , due to ( [ eq : taudt ] ) . for @xmath114
, i.e. for @xmath115 , eq . ( [ 38 ] ) implies that @xmath116 increases faster than in classic diffusion for @xmath117 .
this behavior is an unphysical artifact of the assumption ( [ eq : taudt ] ) . in this regime
the rarefaction rate @xmath112 exceeds the scattering rate @xmath113 , so that scattering can not maintain local thermal equilibrium .
our hydrodynamic diffusion description is therefore not applicable
one must turn to transport theory , as discussed in @xcite .
one can see this explicitly by solving ( [ 25 ] ) for @xmath118 .
linear perturbation analysis reveals that the solutions are unstable for @xmath119 .
we emphasize that for constant coefficients the solutions ( [ 30 ] ) and ( [ 28 ] ) are physical for all @xmath120 , as generally holds for @xmath121 for any @xmath122 . ) , ( [ 38 ] ) and ( [ eq : classict ] ) respectively .
causal curves are for @xmath123 , 1.5 , and 2 .
rapidity spreads are divided by the value @xmath124.,width=288 ] the width computed with varying coefficients increases without bound , albeit slowly .
this behavior distinguishes ( [ 37 ] ) and ( [ 38 ] ) from the constant - coefficient widths obtained from ( [ 30 ] ) and ( [ 28 ] )
. nevertheless , the short time behavior shown in fig .
[ fig:3 ] resembles the constant - coefficient case in fig .
[ fig:1 ] . in both cases
the classic and causal results converge over longer times .
note that the results in fig .
[ fig:3 ] are normalized to @xmath125 using the value of @xmath7 at @xmath66 . to apply these results to collisions
, we must specify the diffusion coefficient @xmath7 and the relaxation time @xmath8 .
transport coefficients in quark gluon plasma and hadron matter have been been studied extensively @xcite .
flavor and charge diffusion estimates in a plasma yield @xmath126 fm and @xmath127 fm @xcite .
the hadron gas case , which is more relevant to this work , is complicated by the menagerie of resonances produced in collisions . in ref .
@xcite , the charge diffusion coefficient was computed using a hadronic transport model , yielding @xmath128 fm and @xmath129 fm .
we argue in sec . [
sec : discussion ] that the short time behavior in figs .
[ fig:1 ] and [ fig:3 ] describes diffusion following hadronization .
hadrons form at a rather late time , roughly @xmath130 fm .
if freeze out occurs shortly thereafter , at @xmath131 less than 20 fm , then the range @xmath132 is important . a hadronic relaxation time
@xmath133 fm is relevant , so that @xmath134 . in this case
causal diffusion gives a much slower spread in rapidity than classic diffusion .
let us now turn to net charge fluctuations and their dissipation . to begin , we review key features of multiplicity and charge fluctuation observables .
our discussion builds on the detailed treatment in ref .
@xcite , which stresses the relation of fluctuation observables to two - particle correlation functions .
we extend that treatment by identifying the net - charge correlation function ( [ eq : q ] ) that drives dynamic charge fluctuations . in the next section
we will show how diffusion affects the evolution of this correlation function .
dynamic fluctuations are generally determined from the measured fluctuations by subtracting the statistical value expected , e.g. , in equilibrium @xcite .
dynamic multiplicity fluctuations are characterized by @xmath135 where @xmath136 here is the event average .
this quantity is obtained from the multiplicity variance by subtracting its poisson value @xmath137 .
similarly , the covariance for different species is @xmath138 which also vanishes for poisson statistics .
these quantities depend only on the two - body correlation function @xmath139 where @xmath140 is the rapidity density of particle pairs for species @xmath141 and @xmath142 at the respective pseudorapidities @xmath143 and @xmath144 , and @xmath145 is the single particle rapidity density .
we use the same symbol for momentum and configuration space rapidity , because we will take these quantities to be equal in later sections . in @xcite
it is shown that @xmath146 these quantities are robust in that they are normalized to minimize the effect of experimental efficiency and acceptance . the star and ceres experiments characterize dynamic net - charge fluctuations using the robust variance @xmath147 proposed in @xcite
; here we use @xmath148 rather than the more conventional notation @xmath149 for simplicity . to establish the relation of @xmath148 to the net charge correlation function @xmath150
we integrate over a rapidity interval to find @xmath151 where the variance of the net charge is @xmath152 and the average number of charged particles is @xmath153 . expanding ( [ eq : omega ] ) yields @xmath154 where @xmath155 .
if the average numbers of particles and antiparticles are nearly equal , @xmath156 hijing simulations show that this relation is essentially exact for mesons at sps and rhic energy ( we find @xmath148 and @xmath157 are equal with 2% statistical error at sps energy ) .
we mention that @xmath158 itself was suggested as an observable in ref .
@xcite and that phenix measures the related quantity @xmath159 .
however , @xmath158 is not strictly robust .
the next section implies that @xmath160 and , consequently , @xmath158 is of more fundamental interest than @xmath148 .
however , in view of ( [ eq : nuomega ] ) and the qualitative aims of fluctuation studies , we will regard these quantities as interchangeable .
to understand how diffusion can dissipate dynamic fluctuations of the net charge and other conserved quantities , we apply the theoretical framework developed by van kampen and others @xcite . a relativistic extension of these techniques will enable the computation of the correlation function ( [ eq : q ] ) as well as statistical quantities like @xmath148 , while introducing no additional parameters .
we start with a single charged species of conserved density @xmath161 that evolves by diffusion .
consider an ensemble of events in which @xmath161 is produced with different values at each point with probability @xmath162 .
after ref .
@xcite , one writes a master equation describing the rate of change of @xmath163 in terms of transition probabilities to states of differing @xmath164 on a discrete spatial lattice .
diffusion and flow are described as transitions in which particles `` hop '' to and from neighboring points . the fokker - planck formulation in ref .
@xcite can be obtained from this master equation in the appropriate limit .
alternatively , one can use the master equation to obtain a partial differential equation for the density correlation function by taking moments of @xmath163 .
the derivation is standard and we do not reproduce it here @xcite . if the one body density follows classic diffusion dynamics , then the density correlation function @xmath165 satisfies the classic diffusion equation @xmath166 @xcite .
this density correlation function is analogous to rapidity - density correlation function ( [ eq : corrfunexp ] ) .
if the event - averaged single - particle density satisfies ( [ eq : grad ] ) , then we can extend this result to causal diffusion by fourier transforming ( [ eq : vardiff ] ) and using ( [ eq : diffomega ] ) , as in sec .
[ sec : diffusion ] .
the inverse transform yields @xmath167 we remark that the delta - function term in ( [ eq : variance ] ) implies that the volume integral of @xmath168 vanishes when particle number fluctuations obey poisson statistics .
this term is derived along with ( [ eq : vardiff ] ) in ref .
@xcite ; it ensures that correlations vanish in global equilibrium . in a nuclear collision with several charged species ,
only the net charge density @xmath169 satisfies diffusion dynamics , since chemical reactions can change one species into another .
we define the net charge correlation function @xmath170 as @xmath171 observe that the integral @xmath172 gives the net charge fluctuations minus its value for uncorrelated poisson statistics .
this integral vanishes in equilibrium . we expand ( [ eq : cgflc0 ] ) to write @xmath173 , where each @xmath174 and @xmath175 has the form ( [ eq : variance ] ) and @xmath176 . to obtain evolution equations for the matter produced in ion collisions ,
we incorporate longitudinal expansion following sec .
[ sec : ions ] .
the net - charge correlation function @xmath177 is given by ( [ eq : q ] ) , where we now take @xmath178 this correlation function has the same form as ( [ eq : cgflc0 ] ) with densities @xmath164 replaced by rapidity densities @xmath179 and @xmath180 replaced by spatial rapidities @xmath181 .
we find that @xmath177 obeys [ 40 ] ( _ d + -d ( _ 1 ^2 + _ 2 ^2 ) ) q=0 for @xmath182 . to compute the effect of diffusion on net charge fluctuations in the next section , we write ( [ 40 ] ) in terms of the relative rapidity @xmath183 and average rapidity @xmath184 : ( _ d + - 2d^2 -d^2 ) q = 0 ; [ 41 ] the `` 2 '' follows from the transformation to relative rapidity @xmath185 . to compute the widths of @xmath186 in relative or average rapidity , one multiplies ( [ 41 ] ) by @xmath187 or @xmath188 and integrates over both variables .
we find @xmath189 and @xmath190 where @xmath191 is calculated using ( [ 28 ] ) ( or ( [ 38 ] ) if a time - dependent @xmath8 and @xmath7 is more appropriate ) .
in this section we discuss how diffusion can affect our ability to interpret hadronization signals . as a concrete illustration , we suppose that collisions form a quark gluon plasma that hadronizes at a time @xmath66 , producing anomalous dynamic charge fluctuations @xmath148 .
the hadronization model of jeon and koch @xcite implies that plasma produces a value @xmath192 .
in contrast , the value for a hadronic resonance gas is variously estimated in the range from @xmath193 @xcite to @xmath194 @xcite .
hadronic diffusion from @xmath66 to a freeze out time @xmath131 can dissipate these fluctuations , reducing @xmath195 .
we ask whether hadronic diffusion can plausibly bring plasma fluctuations near the hadron gas values . while the estimates of ref .
@xcite can be questioned @xcite , they will serve here as benchmarks . in this section
we emphasize the impact of causal diffusion on this problem , employing a simplified approach tailored to that aim .
phenomenological conclusions require a more realistic model . before proceeding
, we emphasize that ref .
@xcite compares quark gluon plasma and hadron gas fluctuations as they might emerge from distinct systems at equivalent chemical potential and temperature .
in contrast , the fluctuations in a hadron gas that has evolved _ from _ a plasma initial state are fixed by charge conservation .
suppose that hadronization in single event produces a positive charge in one rapidity subinterval with a compensating negative charge in another .
the final hadronic state maintains that situation freezing in the plasma fluctuations unless diffusion can redistribute the charges between those subintervals . to @xmath196 ( bottom ) and @xmath197 to @xmath198 ( top).,width=307 ] to compute fluctuation observables following secs .
[ sec : ions ] and [ sec : fluctdiff ] , we specify the correlation function by identifying spatial and momentum - space rapidity at a fixed freeze out proper time @xmath131 .
observe that isr and fnal data @xcite can be characterized as gaussian near midrapidity .
moreover , these data show that charged particle correlations are functions of the relative rapidity @xmath199 with only a weak dependence on the average rapidity @xmath200 . near midrapidity
, these data inspire the form @xmath201 for @xmath202 .
diffusion increases the widths @xmath203 and @xmath204 compared to their initial values @xmath205 and @xmath206 at the hadronization time @xmath66 in accord with ( [ eq : widthrel ] ) and ( [ eq : widthavg ] ) . for simplicity , we assume that @xmath206 is sufficiently large that we can neglect the time dependence of @xmath204 in ( [ eq : widthavg ] ) .
we point out that ( [ eq : qyy ] ) is an exact solution of the classic diffusion equation on an infinite rapidity interval for gaussian initial conditions .
moreover , ( [ eq : qyy ] ) is a good approximation for causal diffusion , provided that the rapidity region of interest does not appreciably exceed @xmath203 or @xmath204 ; see the discussion in sec .
[ sec : diffusion ] .
dynamic fluctuations @xmath148 are computed by integrating @xmath160 over an interval @xmath207 corresponding to the experimental acceptance . for @xmath204 greater than @xmath208 and @xmath203 ,
we use ( [ eq : omega ] ) and ( [ eq : nuomega ] ) to estimate @xmath148 @xmath209 where the total number of charged particles is @xmath210 .
we take the rapidity density @xmath63 for each charge species to be uniform .
the quantity @xmath211 is the value obtained for a large rapidity window .
the dependence of the dynamic fluctuations @xmath212 on the relative rapidity interval @xmath213 implied by ( [ eq : rapdep ] ) is shown in fig .
[ fig : rapidity ] for several ad hoc values of @xmath203 . also shown as dashed curves are the ratios of these quantities , @xmath214 computed for @xmath205 the smaller initial width and @xmath215 the larger final width .
this ratio indicates how much the fluctuations are reduced as the width increases from @xmath205 to @xmath215 . in both figures ,
the change in widths are chosen so that they nearly `` hide '' initial qgp fluctuations at the level expected in ref .
@xcite , since @xmath216 .
balance function measurements are consistent with a width @xmath217 @xcite .
isr and fnal experiments suggest @xmath218 in pp collisions . for the acceptance @xmath219 and @xmath197 .
the gray band indicates the level needed to obscure plasma signals .
solid and dashed curves are respectively computed for constant and varying coefficients using ( [ 28 ] ) and ( [ 38 ] ) and the corresponding classic eqs .
( [ 30 ] ) and ( [ eq : classict]).,width=307 ] the question then becomes : are such increases in @xmath203 plausible in a diffusion model ?
equation ( [ eq : widthrel ] ) implies that @xmath220 .
the relevant `` formation time '' is the time at which hadronization occurs .
this occurs quite late in the evolution , roughly from @xmath221 to 12 fm .
freeze out occurs later , perhaps as late as @xmath222 fm .
the ratio @xmath223 is then in the range from one to two , where the difference between causal and classic diffusion is substantial , as shown in figs .
[ fig:1 ] and [ fig:3 ] .
we take @xmath224 fm and @xmath129 fm from ref .
@xcite , as discussed in sec .
[ sec : ions ] . the solid curves in fig .
[ fig : plasma ] show the decrease of dynamic fluctuations computed using ( [ eq : rat ] ) , ( [ eq : widthrel ] ) , and ( [ 28 ] ) .
the dashed curves are computed using ( [ 38 ] ) instead of ( [ 28 ] ) .
we take the value @xmath225 for the initial width ; larger values imply a smaller net reduction of @xmath212 .
a hadronization time @xmath226 fm is assumed , corresponding to @xmath227 in figs .
[ fig:1 ] and [ fig:3 ] .
the gray band indicates the level to which @xmath212 must be reduced to hide the effect of plasma fluctuations .
we see that classic diffusion can bring @xmath212 to this level if the system lives as long as 20 fm , while causal diffusion can not .
in this paper we introduce a causal diffusion equation to describe the relativistic evolution of net charge and other conserved quantities in nuclear collisions . we find that causal limitations inhibit dissipation . to study the effect of this dissipation on fluctuation signals , we obtain a causal diffusion equation for the two - body correlation function .
our approach complements the treatment in @xcite , but is more easily generalized to include radial flow and other dynamic effects for comparison to data .
we then use these equations to provide estimates that show that net charge fluctuations induced by quark gluon plasma hadronization can plausibly survive diffusion in the hadronic stage .
this result bolsters our optimism for fluctuation and correlation probes of hadronization and other interesting dynamics .
our aim here has been to emphasize the impact of causal diffusion on charge fluctuations .
correspondingly , we employ an idealized approach tailored to that aim .
phenomenological conclusions require a more realistic model , to be developed elsewhere .
our results on post - hadronization diffusion owe to the similarity of the lifetime of the hadronic system and the relaxation time for diffusion @xmath8 .
in fact , rhic data suggests hadronic lifetimes that are as short or shorter than we assume : hbt @xcite and resonance @xcite measurements suggest a lifetime from roughly @xmath228 to @xmath229 fm . on the other hand ,
our estimate of @xmath8 rests on transport calculations in ref .
the importance of this problem invites further study of charge and other transport coefficients . in closing , we briefly comment on the experimental situation .
charge fluctuation measurements give roughly similar values from sps to rhic energy @xcite .
furthermore , the star experiment finds @xmath230 at rhic @xcite .
while this value is close to the benchmark hadron gas estimates @xcite , we emphasize that those estimates were constructed for an equilibrium resonance gas assuming no prior plasma stage . flow and jet quenching measurements provide strong indications if not proof of plasma formation @xcite .
as discussed earlier , @xmath212 can be very different in a resonance gas formed from a hadronizing partonic system and , indeed , should be very close to plasma values .
extraordinary charge fluctuations should be seen , but are not .
it may be that net - charge fluctuations produced by plasma hadronization are closer to the benchmark hadronic estimates , as suggested by bialas @xcite .
what then ?
experimentally , one can turn to the centrality and azimuthal - angle dependence of fluctuations and , eventually , to correlation - function measurements as more sensitive probes of net charge fluctuations .
such studies have already begun to yield important information @xcite .
in addition , one can study baryon number and isospin fluctuations , which are more closely related to the qcd order parameter and , therefore , to hadronization physics and phase transition dynamics @xcite .
we thank r. bellwied , h. denman , l. d. favro , a. muronga , c. pruneau , and d. teaney for discussions .
this work was supported in part by a u.s .
national science foundation career award under grant phy-0348559 .
m. a. a. was supported by a thomas c. rumble university graduate fellowship at wayne state university .
99 j. t. mitchell [ phenix ] , proc .
quark matter 2004 , to be published j. phys .
g. ( 2004 ) arxive : nucl - ex/0404005 ; g. westfall [ star ] _ ibid _ , arxive : nucl - ex/0404004 ; c. pruneau [ star ] , arxiv : nucl - ex/0401016 .
g. baym , g. friedman and i. sarcevic , phys .
* b219 * , ( 1989 ) 205 ; m. gazdzicki and s. mrowczynski , z. phys .
* c54 * ( 1992 ) 127 ; h. heiselberg , phys .
* 351 * ( 2001 ) 161 ; s. y. jeon and v. koch , arxiv : hep - ph/0304012 . m. stephanov , k. rajagopal and e. v. shuryak , phys . rev .
d * 60 * , 114028 ( 1999 ) [ hep - ph/9903292 ] .
m. asakawa , u. heinz , b. mller , phys .
rev . lett . * 85 * , 2072 ( 2000 ) ; s. jeon and v. koch,_ibid _ , 2076 ; v. koch , m. bleicher , s. jeon , nucl . phys .
* a698 * ( 2002 ) 261 .
d. bower and s. gavin , phys . rev .
c * 64 * 051902 ( 2001 ) ; y. hatta and m. a. stephanov , phys .
lett.*91 * , 102003 ( 2003 ) ; erratum , _ ibid _ * 91 * 129901 ( 2003 ) .
e. v. shuryak and m. a. stephanov , phys .
c * 63 * , 064903 ( 2001 ) [ arxiv : hep - ph/0010100 ] .
a. muronga , phys .
* 88 * , 062302 ( 2002 ) ; phys .
c * 69 * , 034903 ( 2004 ) .
d. teaney , arxiv : nucl - th/0403053 .
m. abdel aziz and s. gavin , proc .
19th winter workshop on nuclear dynamics , breckenridge , colorado , usa , february 8 - 15 , 2003 , ( ep systema bt . , debrecen , hungary , 2003 ) , 159 . n. g. van kampen , _
stochastic processes in physics and chemistry _ , ( elsevier science , amsterdam , 1997 ) ; c. w. gardiner , _ handbook of stochastic methods for physics , chemistry and the natural sciences _ , ( springer , berlin , 2002 ) .
c. cataneo , atti semin . mat .
modena * 3 * , 3 ( 1948 ) ; p. vernotte , c. r. acad .
paris * 246 * , 3154 ( 1958 ) .
h. grad , comm .
pure appl .
math . 2 , 331 ( 1949 ) .
w. israel and j. m. stewart , ann .
118 , 341 ( 1979 ) .
e. zauderer , _ partial differential equations of applied mathematics _ , 2nd ed .
( wiley , new york , 1989 ) .
d. d. joseph and l. preziosi , rev .
61 * , 41 ( 1989 ) .
s. gavin , nucl .
a * 435 * , 826 ( 1985 ) .
d. m. forster , _ hydrodynamic fluctuations , broken symmetry , and correlation functions _ , ( benjamin , new york , 1975 ) . c. manuel and s. mrowczynski , arxiv : hep - ph/0403054 .
m. prakash , m. prakash , r. venugopalan and g. welke phys .
rept . * 227 * , 321 ( 1993 ) .
a. muronga , phys .
c * 69 * , 044901 ( 2004 ) .
h. heiselberg , c.j .
pethick phys .
d48 , 2916 , ( 1993 ) .
peter arnold , guy d. moore , and lawrence g. yaffe , jhep * 0305 * , 051 , ( 2003 ) [ arxiv : hep - ph/0302165 ] .
n. sasaki , o. miyamura , s. muroya and c. nonaka , europhys .
lett . , 38 ( 2001 ) [ arxiv : hep - ph/0007121 ] .
s. r. de groot , h. a. van leeuven and c. g. van weert , _ relativistic kinetic theory _ , ( north holland , amesterdam , 1980 ) .
j. d. bjorken , phys .
d * 27 * , 140 ( 1983 ) .
s. gavin , nucl .
b * 351 * , 561 ( 1991 ) . c. pruneau , s. gavin and s. voloshin , phys .
c * 66 * , 044904 ( 2002 ) [ arxiv : nucl - ex/0204011 ] .
j. adams et al .
[ star ] , phys .
c bf 68 , 044905 ( 2003 ) .
a. bialas , phys . lett .
b * 532 * , 249 ( 2002 ) [ arxiv : hep - ph/0203047 ] .
j. whitmore , phys .
* 27 * , 187 ( 1976 ) ; l. foa , physics reports , 22 , 1 ( 1075 ) ; h. boggild and t. ferbel , ann .
* 24 * , 451 ( l974 ) .
s. a. bass , p. danielewicz and s. pratt , phys .
lett . *
85 * , 2689 ( 2000 ) [ arxiv : nucl - th/0005044 ] .
m. lisa [ star ] , acta phys .
* b35 * , 37 ( 2004 ) . c. markert [ star ] , arxiv : nucl - ex/0404003 .
m. gyulassy , arxiv : nucl - th/0403032 .
s. gavin , phys .
, to be published ( 2004 ) [ arxiv : nucl - th/0308067 ] ; j. phys .
g. , to be published ; [ arxiv : nucl - th/0404048 ] . | diffusion may obliterate fluctuation signals of the qcd phase transition in nuclear collisions at sps and rhic energies . we propose a hyperbolic diffusion equation to study the dissipation of net charge fluctuations .
this equation is needed in a relativistic context , because the classic parabolic diffusion equation violates causality .
we find that causality substantially limits the extent to which diffusion can dissipate these fluctuations . |
the munu experiment was designed to study @xmath1 scattering at low energy , probing in particular the magnetic moment of the neutrino .
the detector was set up near a nuclear power reactor in bugey ( france ) serving as antineutrino source .
we report here on an analysis of data corresponding to 66.6 days of live time reactor - on and 16.7 days reactor - off .
many experiments now show that neutrinos have masses , and that the weak eigenstates @xmath2 are superpositions of the mass eigenstates @xmath3 : @xmath4 intense work is being done to determine more precisely the mixings @xmath5 and the masses @xmath6 .
besides masses , neutrinos can have magnetic moments . as shown in ref .
@xcite , the fundamental magnetic moments are associated with the mass eigenstates , in the basis of which they are represented by a matrix @xmath7 ( @xmath8 ) .
dirac neutrinos can have both diagonal and off - diagonal ( transition ) moments , while off diagonal moments only are possible with majorana neutrinos .
astrophysical considerations put strong constraints on dirac neutrinos , and less strong ones on majorana neutrinos ( see @xcite ) .
upper limits applying in both cases from stellar cooling of order @xmath9 have been derived .
magnetic moments large enough would lead to spin - flavor precession in the toroidal magnetic field in the interior of the sun , as discussed for instance in @xcite , further complicating the oscillation pattern of solar neutrinos resulting from masses and mixings .
these astrophysical bounds on magnetic moments , however , depend on various assumptions and , to a certain extent , are model dependent .
direct measurements are under much better control .
so far the best limits stem from experiments studying @xmath10 or @xmath0 scattering , and looking for deviations of the measured cross - section from the one expected with weak interaction alone .
the cross - section is given by @xmath11 \nonumber\\ & & \mbox { } + \frac{\pi \alpha^{2}\mu_{e}^{2}}{m_{e}^{2}}\frac{1-t / e_{\nu } } { t } \end{aligned}\ ] ] with the contribution of the weak interaction in the first two lines , and that from the magnetic moment @xmath12 in the last one @xcite . here
@xmath13 is the incident neutrino energy , @xmath14 the electron recoil energy , and the couplings are given by @xmath15 the relative contribution of the magnetic moment term increases with decreasing neutrino and electron energies .
therefore it is essential to have a low electron detection threshold , looking for neutrinos from a low energy source .
so far the sun and nuclear reactors have been used .
future experiments with radioactive sources are planned .
the measured squared magnetic moment @xmath16 depends on the mixings and , in case of a large distance @xmath17 from the source to the detector , on the propagation properties of neutrinos . for vacuum oscillations it is given by : @xmath18 with @xmath19 the momentum of @xmath20 with mass @xmath21 , for a given neutrino energy @xmath13 . in the case of matter
enhanced oscillations the dependence of the propagation eigenstates on the local density must be taken into account @xcite . here
the source to detector distance is long , oscillations are relevant and the measured quantity is denoted @xmath22 .
super - kamiokande has measured with a very high statistical accuracy the electron recoil spectrum from the scattering of solar @xmath23b neutrinos @xcite .
the total rate is reduced because of neutrino oscillations , but the shape is seen to be in good agreement , within statistics , with that expected assuming weak interaction alone .
beacom and vogel @xcite looked to what extent excess counts at the low energy end from a magnetic moment can be ruled out .
the assumption is made that such an excess is not compensated by a distortion due to oscillations .
the limit @xmath24 @xmath25 at 90 % confidence level ( cl ) was derived .
nuclear reactors are strong sources of @xmath26 with energies ranging up to about 8 mev .
these are essentially produced in the beta decay of fission fragments , with however a significant contribution at low energy from nuclei activated by neutrons . above 1.5 - 2.0 mev
the integral beta spectrum of the fission fragments of the isotopes in which fission is predominantly induced by thermal neutrons ( @xmath27u , @xmath28pu , @xmath29pu ) was measured .
these isotopes dominate .
a simple procedure leads to the corresponding neutrino spectra , which are thus known well , with a precision of order 5 % or better @xcite . in @xmath30u ,
fission is induced by fast neutrons , and no such measurements are available .
less precise calculations ( @xmath3110 % ) only can be used .
but this isotope contributes only about 6 - 7 % in a conventional reactor .
the sum spectrum above 2 mev can be reconstructed with a precision of order @xmath325 % , knowing the relative contributions of the various fissile isotopes at a given reactor .
studies of @xmath33 scattering at reactors @xcite , which sample the antineutrino spectrum above 2.5 mev , show good agreement with these predictions .
below 1.5 - 2 mev the neutrino spectrum can only be reconstructed from inherently less precise calculations as shown in ref .
@xcite . in that work
it is mentioned that neutron activations of the fissile isotopes , leading in particular to @xmath28u , @xmath28np and @xmath34u , and of the fission products , contribute .
the yields have been estimated . in reactor experiments
the source to detector distance is short compared to oscillation lengths , and the magnetic moment searched for @xmath35 is given by eq .
( [ mudef ] ) with l set to zero .
the irvine group was the first to observe @xmath0 scattering @xcite .
the experiment was performed at the savannah river reactor , with a 16 kg plastic scintillation counter surrounded by a nai veto at 11 m from the reactor core .
a reactor - on minus reactor - off signal was seen , with a threshold at 1.5 mev . analyzing the electron recoil data with the most recent knowledge of the weinberg angle and the reactor spectrum , vogel and
engel @xcite found a slight excess of events which can be explained by a magnetic moment @xmath35 of order @xmath36 @xmath25 .
also @xmath0 scattering was studied at the rovno reactor by a group from saint petersburg .
the detector consisted of a stack of silicon sensors , with a total mass of 75 kg .
a reactor - on minus reactor - off signal was seen above 600 kev , with a signal to background ratio of order 1:100 .
the limit @xmath37 @xmath25 at 95(68 ) % cl was reported @xcite .
more recently the texono collaboration installed an ultra low background high purity germanium detector , with a fiducial mass of 1.06 kg , near the kuo - sheng reactor in taiwan @xcite . here
the approach is somewhat different .
the threshold on the electron recoil ( 12 kev ) is extremely low .
the reactor - on and reactor - off spectra were found to be identical within statistical errors . from
that the limit @xmath38 @xmath25 at 90(68 ) % cl was derived .
in the aforementioned experiments the energy of the recoil electron candidate only was measured .
the munu collaboration @xcite has built a detector of a different kind , in which the topology of events is recorded .
this allows a better event selection , leading to a lower background .
moreover , in addition to the energy , the initial direction of an electron track can be measured .
a second parameter , the electron scattering angle , can therefore be reconstructed .
this allows to look for a reactor signal by comparing forward electrons , having as reference the reactor to detector axis , with the backward ones .
the background is measured on - line , which eliminates problems from detector instabilities , as well as from a possible time dependence of the background itself .
the detector is described in details in ref .
@xcite , and we only present here the essential features .
the detector , made from radiochemically clean materials , was installed at 18 m from the core of a commercial reactor in bugey ( france ) with a power of 2750 mwth .
it emits neutrinos from the fission fragments of @xmath27u ( 54 % on average over an annual reactor cycle ) , @xmath28pu ( 33 % ) , @xmath29pu ( 6 % ) , and @xmath30u ( 7 % ) .
the central component of the munu detector consists of a time projection chamber ( tpc ) filled with 3 bar of cf@xmath39 gas ( figure [ fi : munu0 ] ) , acting as target and detector medium for the recoil electron .
cf@xmath39 was chosen because of its relatively low z , leading to reduced multiple scattering , its absence of protons , which eliminates backgrounds from @xmath40 scattering , its good drifting properties and its high density .
as shown in figure [ fi : munu0 ] the gas is contained in a cylindrical acrylic vessel of 1 m@xmath41 volume ( 90 cm in diameter , 162 cm long , total cf@xmath39 mass 11.4 kg ) , which is entirely active .
the drift field parallel to the tpc axis ( z - axis ) is defined by a cathode on one side , a grid on the other side , and field shaping rings at successive potentials outside the acrylic vessel .
the voltages were set to provide a homogeneous field leading to a drift velocity of 2.14 @xmath42s@xmath43 . an anode plane with 20 @xmath44 m wires and a pitch of 4.95 mm , separated by 100 @xmath44 m potential wires ,
is placed behind the grid , to amplify the ionization charge .
the integrated anode signal gives the total energy deposit .
a pick - up plane with perpendicular x and y strips ( pitch 3.5 mm ) behind the anode provides the spatial information in the x - y plane perpendicular to the z - axis .
the spatial information along the z - axis is obtained from the time evolution of the signal . to reduce systematics
when comparing forward and backward events , the tpc was positioned orthogonally to the reactor - detector axis , which moreover coincides with the bisecting line between x and y strips on the pick - up plane .
the anode wires are rotated by 45@xmath45 with respect to the x - y plane .
the tpc is thus absolutely symmetric between backward and forward directions with regards to the reactor - detector axis .
the imaging capability of the tpc is illustrated in figure [ fi : munu_track ] which shows an electron track .
[ fi : munu_track ] the energy resolution and calibration is obtained by comparing spectra measured using various @xmath46 sources ( @xmath47cs ( 662 kev ) , @xmath48mn ( 835 kev ) and @xmath49na ( 1274 kev ) ) with simulations .
the relative energy resolution is found to be 8 % ( 1 @xmath50 ) at 1 mev , correcting for small variations across the anode plane .
it scales with the power 0.7 of the energy , rather than with the square root .
we think that this is because the electron attachement in cf@xmath39 in the strong field around the anode wires is rather high , affecting the statistics in the avalanche @xcite .
the gain stability is monitored regularly throughout the data taking period , with sources and by measuring the spectrum of cosmic muons crossing the tpc .
the data are corrected for small instabilities .
the angular resolution is around 10@xmath45 ( 1 @xmath50 ) at 1 mev , for tracks scanned visually , as derived from monte - carlo simulations , with a slight angular dependence @xcite .
the energy dependence is only weak above 700 kev .
the acrylic vessel is immersed in a steel tank filled with liquid scintillator and viewed by photomultipliers , 24 on each side .
it acts as a veto counter against the cosmics and as an anti - compton detector to reduce the background from @xmath46 s .
the scintillator also sees the light produced by the avalanche around the anode wire , providing an additional measurement of the ionisation charge .
the primary light of heavily ionizing particles such as @xmath51 s is seen as well .
a few @xmath51 s are observed from remaining surface contaminations on the cathode .
the primary light of minimum ionizing particles confined in the gas volume however is below detection threshold .
the tpc threshold is set at 300 kev , and that in the scintillator at 100 kev , with a minimum of 5 photomultipliers hit .
we note that the singles rates in the photomultipliers show no azimuthal dependence , which could result from hot spots . in normal data taking tpc events above threshold
are read out , provided they are not in coincidence with a 200 @xmath44s signal started by scintillator pulses above 22 mev .
this eliminates direct cosmic hits , or neutrons associated with them .
the 48 photomultipliers are read out first .
the light from the photomultipliers on the anode side and the cathode side is compared .
reading proceeds only if the relative difference is less than @xmath3230 % . events corresponding to real tracks inside the gas volume were found to always fall within these limits @xcite .
discharges however can result in a larger imbalance , in which case they are readily eliminated .
this helps reducing the average readout time .
good events are single electrons contained in the tpc volume .
the selection of neutrino scattering events proceeds in two steps .
first an automatic filtering eliminates obviously bad events , namely events * identified as @xmath51 s or discharges from their topology ( high ionization in a small volume ) , * or in delayed coincidence ( 80 @xmath44s , corresponding to the tpc length ) with a signal in the scintillator , * or events not contained in a fiducial volume of 42 cm radius , * and finally events with a fast rise time , due to particles crossing the amplification gap between the pick up plane plane and the grid , in either direction . here
the avalanche light signal is used , as described in ref .
@xcite . the precise live time is derived in this process on a daily basis .
it is found to fluctuate around 65 % .
it is limited primarily by the total veto time of the scintillator ( 11 % ) and the dead time of the tpc itself , caused by the relatively long data read - out and data transfer time ( 24 % ) . then a final scan is performed . in a separate publication , we reported first results from a crude automatic procedure @xcite , carried out with a pattern recognition program . here
we present an analysis based on a visual scan of events , which has the advantage of cleaner event selection .
also the risk of misidentifying the beginning of a track is much reduced .
this approach is time consuming , however . to minimize the work load , that analysis was restricted to energies above 700 kev .
results from an improved version of the automatic scanning procedure will be discussed in a subsequent section .
the data sets corresponds to 66.6 days of live time reactor - on and 16.7 days reactor - off . in this second scan both the @xmath52 and @xmath53 projections of an event
are scrutinized , as well as the evolution in time of the anode and scintillator signals .
continuous electron tracks only are retained .
the end of the track is identified from the increased energy deposition , due to the higher stopping power , as exemplified in figure [ fi : munu_track ] .
events with a second high deposition along the track , in particular near the other end , are discarded .
this improves the background suppression , at the cost of events with a delta electron , which however contribute negligibly .
then the tangent at the start of the track is determined by eye . from that the angles @xmath54 with respect to the reactor core - detector axis is determined , as well as the angle @xmath55 with respect to the tpc axis . an excess of events from the anode side was observed .
it is presumably due to additional activities , resulting from the greater complexity of the readout system , and to a larger inactive volume in the scintillator because of the stronger and thicker acrylic lid .
for that reason only electrons emitted in the half sphere from the cathode side ( @xmath56 ) are accepted .
this reduces the acceptance by a factor 2 , but leads to a better signal to background ratio .
in figure [ fi : theta_ron ] we show the distribution of @xmath57 of single contained electrons , for both reactor - on and reactor - off .
the slightly non linear angular response of the tpc , and its geometry , explain the accumulation of events at @xmath57 around 1 , -1 and 0 for reactor - off .
the distribution is however identical in forward ( @xmath58 ) and backward ( @xmath59 ) directions .
the reactor - on spectrum shows a clear excess of events in forward direction from @xmath0 scattering .
uncertainties from instabilities of the gain , the veto rate , the live time , cancel out in this forward minus backward comparison . for each electron event
the neutrino energy @xmath13 is reconstructed from the electron recoil energy @xmath14 and from the scattering angle , taken as @xmath54 .
forward event candidates are defined as those with positive neutrino energy : @xmath60 . to select backward events
the neutrino energy which must be positive is that calculated using @xmath61 as scattering angle .
this procedure has an acceptance close to 100 % above 700 kev , as determined by monte - carlo simulations .
it takes into account that electrons are emitted in an narrower cone whith increasing energies , and is somewhat more sophisticated than the application of a crude cut on @xmath57 . in practice , however , with a threshold of 700 kev , it produces almost the same result as the application of a cut @xmath62 or @xmath63 .
the energy distributions of both forward and backward events are displayed in figure [ fi : fbe ] . a clear excess of forward events ( 458 in total ) over backward events ( 340 ) is seen .
the total forward minus backward count rate above 700 kev is thus [email protected] day@xmath43 .
the background , given by the backward events , shows a steep low energy component ending at about 1.2 mev . above it
is seen to be fairly low , and much flatter .
the background was observed to increase slightly during the course of the experiment , by about 10 % in 6 months , possibly because of outgassing . a small drop in efficiency of the anti - compton , even without an observable reduction in count rate , could also explain this . as a cross check the reactor - off recoil spectra , both forward and backward , measured after the reactor - on period ,
were reconstructed as well , and are displayed in figure [ fi : fbeoff ] .
they are indeed identical within statistics , the integrated forward minus backward rate above 700 kev being @xmath64 day@xmath43 .
the difference of the forward minus backward reactor - on spectra is shown in figure [ fi : fminusbe ] .
the expected event rate was calculated using the best knowledge of the reactor spectrum as described above , and taking into account the known activations of fissile isotopes and fission products .
the uncertainty is around 5 % above 900 kev , as discussed previously , and larger below .
the various acceptances of the event selection procedure were determined from measurements with sources , and by monte carlo simulations using the geant3 code .
data sets including recoil electrons from neutrino scattering were produced , filtered and scanned just as the real data .
the containment efficiency in the 42 cm fiducial radius for recoil electrons was found to vary from 63 % at 700 kev , 50 % at 1 mev to 12 % at 2 mev .
the relative uncertainty is of order 2 - 3 % .
some tracks with weird topologies can not be reconstructed , reducing the acceptance by 4 % , relatively speaking . similarly the remaining acceptances together , including that of the @xmath60 cut , lead to an additional 6 % reduction .
neutrino interactions in the copper cathode give rise to electrons which can escape into the gas volume .
these can not be vetoed .
taking into account all cuts they increase the expected rate between 700 kev to 2 mev by some 3 % .
the relative uncertainty on the global acceptance is of order 7 % , leading to a total uncertainty of 9 % on the total expected rate .
the total expected rate above 700 kev assuming a vanishing magnetic moment was found to be [email protected] day@xmath43 , to be compared with [email protected] day@xmath43 measured , as mentioned above .
there is thus a certain excess of measured events , which has however only a small statistical significance .
the calculated energy distribution is shown in figure [ fi : fminusbe ] .
the excess counts are seen to be in the region below 900 kev . to be more quantitative ,
a @xmath65 was calculated using 100 kev bins as shown in figure [ fi : fminusbe ] from 700 kev to 1400 kev , and then a bin from 1400 to 2000 kev .
gaussian statistics applies then to all bins .
the error on the expected rate is small in comparison to the statistical uncertainties , and turns out to be negligible .
the @xmath65 was calculated this way for several values of the squared magnetic moment @xmath66 , constraining it to the region @xmath67 @xmath68 .
this is consistent with a vanishing magnetic moment .
renormalizing to the physical region @xmath69 we find the limit @xmath70 @xmath71 at 90(68)% cl .
nevertheless even the best fit is not very satisfactory . the @xmath65 ( 9.9 for 7 degrees of freedom )
is on the high side .
more troublesome , as seen in figure [ fi : fminusbe ] , the inclusion of a magnetic moment in the calculated spectrum improves the agreement with the data in the two first bins from 700 to 900 kev , but makes it worse in the upper ones .
these two first bins are solely responsible for the high total rate .
we note that such an excess is also visible , unfortunately also with limited statistical precision , and appearing at somewhat smaller energies , in the data of ref .
this excess may well result not from a magnetic moment , but instead from sources not taken into account when evaluating the reactor neutrino spectrum .
as mentioned above , the low energy part , which contributes significantly to the electron spectrum below 0.9 mev , is not so well known .
additional neutron activations may contribute , beyond the known ones of the fissile isotopes and the fission products . in that sense
it seems safe at this stage to restrict the analysis to electron energies above 900 kev .
there the measured event rate is @xmath72 day@xmath43 , in good agreement with the expected one @xmath73 day@xmath43 . from the event rate and the energy distribution , using the same prescription as above , the allowed range @xmath74 @xmath68 is found .
the best @xmath65 is 1.51 for 5 degrees of freedom .
this yields the limit @xmath75 somewhat more stringent than the previous one .
extending the visual scan to energies below 700 kev is not feasible because of the rapid increase of the background below that energy , leading to an unmanageable workload .
but the automatic scanning procedure mentioned in @xcite has been improved , and used on the same data set @xcite .
the program first identifies electron tracks , searches the vertex , and then fits the beginning of the track .
electron tracks as identical as possible to real tracks were produced by monte - carlo .
the electronics noise was included .
the tracks were analyzed with the same program to determine the global acceptance , as well as the angular resolution .
the angular resolution is inferior to that obtained in the visual scanning , varying from 31@xmath45 at 300 kev , to 18@xmath45 at 700 kev and finally 12@xmath45 at 2 mev .
this reduces the acceptance of the @xmath60 cut in particular at low energy .
this , combined with poorer electron track identification , leads to an acceptance of 20 % at 300 kev , increasing to 54 % at 1 mev and remaing constant above that energy .
events with @xmath76 were taken .
the other cuts and acceptances are the same as in the visual scanning .
the background suppression is inferior by a factor 3 or more , so that the statistical precision is less .
but the method can be applied to energies extending down to 300 kev .
we show in figure [ fi : f - be_auto ] the forward minus backward spectra for reactor - on .
it is in good general agreement with the spectra from the eye scan .
the high bin at 750 kev is reproduced .
but , admittedly with relatively large statistical errors , a catastrophic rise when going to lower energies can be ruled out . applying the same procedure as above to the data one
obtains the limit @xmath77 @xmath71 at 90 % cl .
it does not change with the threshold in the range 300 to 800 kev , the better sensitivity at low energy being offset by the larger statistical uncertainties .
the munu experiment studied @xmath0 scattering at low energy near a nuclear reactor . for electron recoil energies above 900 kev
, good agreement is seen with expectations assuming weak interaction alone . from this the following limit on the magnetic moment of the neutrino can be derived : @xmath78 @xmath71 at 90(68 ) % cl , limited primarily by statistics . in any event
this bound rules out the possible indication for a relatively large magnetic moment from the savannah river experiment @xcite , and improves on the limit from ref .
@xcite and @xcite .
it is somewhat more stringent than the limit on @xmath79 from solar neutrinos @xcite , which however does not apply to exactly the same quantity .
99 b. kayser , phys .
d26 ( 1982 ) 1662 r. e. shrock , nucl .
b206 ( 1982 ) 359 j. f. beacom and p. vogel , phys .
83 ( 1999 ) 5222 g. g. raffelt , _ stars as laboratories for fundamental physics _ , university of chicago press , chicago 1996 the munu collaboration ( c. amsler et al . ) , nucl .
inst . and meth . a396 ( 1997 )
akhmedov , a. lanza and s.t .
petcov , phys .
b303 ( 1993 ) 85 e. kh .
akhmedov and j. pulido,_solar neutrino oscillations and bounds on neutrino magnetic moment and solar magnetic field _ , hep - ph/0209192 , 2002 p. vogel and j. engel , phys .
d39 ( 1989 ) 3378 the superkamiokande collaboration ( y. fukuda et al .
82 ( 1999 ) 2430 g. zacek et al .
d34 ( 1986 ) 2621 b. achkar et al . ,
b534 ( 1995 ) 503 v.i .
kopeikin et al
. phys . of atomic nuclei vol . 60
( 1997 ) 172 f. reines , h.s .
gurr and h.w .
sobel , phys .
37(1976)315 a.i .
derbin et al .
, jetp lett . , 57 ( 1993 ) 768 the texono collaboration ( h.b.li et al . ) , phys .
( 2003 ) 131802 the munu collaboration ( m.avenier et al . ) , nucl .
inst . and meth .
a482 ( 2002 ) 408 j. lamblin , thesis , institut national polytechnique de grenoble , 2002 z. daraktchieva , thesis , university of neuchtel , in preparation o. link , thesis , university of zurich , 2003 the munu collaboration ( c. amsler et al . ) , phys .
lett.b 545 ( 2002 ) 57 | the munu experiment was carried out at the bugey nuclear power reactor .
the aim was the study of @xmath0 elastic scattering at low energy .
the recoil electrons were recorded in a gas time projection chamber , immersed in a tank filled with liquid scintillator serving as veto detector , suppressing in particular compton electrons . the measured electron recoil spectrum is presented .
upper limits on the neutrino magnetic moment were derived and are discussed . |
the concept of active brownian motion was developed to describe complex motion in various systems @xcite , @xcite .
its main idea is to assume the existence of an additional _ internal _ degree of freedom , called the internal energy @xmath0 , which can be taken up from the environment and converted to energy of motion .
former models only assumed an exchange between the internal energy and the _ kinetic _ energy of motion . in @xcite , we discussed a more general case , where @xmath0 can be also converted to the potential energy of the particle , thus to the _ full _ mechanical energy @xmath1 .
the corresponding stochastic equations for the position @xmath2 and the momentum @xmath3 of the particle read @xmath4 @xmath5 where @xmath6 represents the stochastic noise , satisfying the correlations @xmath7 and therefore describing a completely irregular force .
the constants @xmath8 and @xmath9 remain to be fixed for concrete applications .
+ + the above system of equations defines the dynamics of a particle subject to an external potential @xmath10 and an irregular force @xmath6 .
the friction with the environment is characterized by the friction constant @xmath11 .
the dynamical equation for @xmath0 has the following meaning : the first term gives the constant rate of internal energy take - up from the environment , the second term @xmath12 models the inner energy loss of the indivual , and the third term @xmath13describes the conversion of internal energy @xmath0 into the mechanical energy @xmath14 .
numerical simulations and analytic investigations of these dynamics were made in @xcite for a single particle in harmonic external potentials .
studies of the multi - particle case were missing so far .
+ + in @xcite many particle systems with active brownian motion were discussed and a specific form of circular swarming behavior was found . in this case , however , internal energy was converted only to _ kinetic _ energy ( corresponding to @xmath15 and @xmath14 being replaced by @xmath16 in the equation for @xmath17 ) .
further simplifications were made by studying only situations , where the internal energy @xmath0 was considered to be stationary ( @xmath18 ) and thus could be eliminated from the dynamical equations .
+ + we now want to discuss in full generality the behavior of many particles undergoing _ canonical _ active brownian motion , i.e. study particles which can convert dynamical internal energy to _ total _ mechanical energy .
we will show amorphous swarming behavior as well as static configurations .
we present a simple model for a swarm of active brownian particles by assuming the existence of a single global internal energy @xmath0 of the swarm . consider the movement of @xmath19 active brownian particles , enumerated by the index @xmath20 , with unit mass in @xmath21 space dimensions coupled to the internal energy @xmath0 .
the position vectors are @xmath22 with @xmath23 and the stochastic equations of motion read @xmath24 @xmath25 @xmath26 where the total mechanical energy is given by @xmath27 the potential energy of one particle is composed of an potential @xmath28 , describing the interaction with all the other particles , and an external potential @xmath29 , modeling the environment of the swarm .
we study the case where all individuals are attracted to the center of mass @xmath30 and are moving in a landscape of harmonic shape .
the total mechanical energy therefore reads @xmath31 ( with positive constants @xmath32 and @xmath33 ) so that the swarm dynamics is finally given by the equations @xmath34 the above system of coupled stochastic nonlinear differential equations is not easily accessible with direct analytic procedures .
nonetheless , predictions for the long time behavior can be made by transforming to the new variables @xmath35 @xmath36 @xmath37 represents the total kinetic energy of the swarm , @xmath38 the total external potential energy , whereas the variables @xmath39 and @xmath40 are quadratic forms of the center of mass @xmath41 , resp . its time derivative @xmath42 : @xmath43 and @xmath44
additionally , also the variable @xmath45 can be written in terms of @xmath41 and @xmath46 : @xmath47 . in these new variables ,
when omitting the noise influences , the system reads : @xmath48 to handle the above nonlinear system , we first calculate stationary points , i.e. values of @xmath49 , where @xmath50 and linearize the equations around these equilibrium solutions . by calculating eigenvalues of the jacobian
, we can analyze stability properties .
we find three distinct equilibrium points @xmath51 : @xmath52 equilibria @xmath53 and @xmath54 are truly _ static _ configurations of the swarm , since @xmath55 in both cases , whereas equilibirum @xmath56 has a nonvanishing equilibrium value of @xmath57 , which means that swarm particles are actually moving .
+ + calculating the stability conditions of all three equilibria allows us to decide under which parameter conditions the swarm either collapses to a point ( @xmath53 ) , freezes to a certain pattern ( @xmath54 ) , or moves amorphously ( @xmath56 ) . for equilibrium @xmath53 ,
all seven eigenvalues of the jacobian are accessible and it is found to be stable under the conditions @xmath58 if these inequalities are satisfied , all swarm particles will collapse into the origin and will remain motionless apart from stochastic fluctuations .
time evolution of the deterministic equations ( [ eins])-([zwei ] ) for @xmath37 ( dashed line ) , @xmath38 ( solid - line ) , @xmath59 ( dot - dashed line ) , @xmath0 ( solid line ) , @xmath45 , @xmath39 and @xmath40 ( not shown for simplicity of presentation ) for the collapse of the swarm to the minimum of the external potential ( @xmath53 ) .
simulation parameters are @xmath60 , @xmath61 , @xmath62 , @xmath63 , @xmath64 and @xmath65 ( @xmath66 ) . ]
the evolution of such a collapse , according to equations ( [ eins])-([zwei ] ) is shown in figure [ fig : one ] .
more interesting is equilibrium @xmath54 , which is stable under the conditions @xmath67 , @xmath68 and @xmath69 we used the routh - hurwitz - theorem @xcite to determine the stability properties .
exactly one eigenvalue is identical to zero independet of the parameter choice .
@xmath54 is in fact not an isolated fixpoint , but a one - dimensional continuum of equilibria , since @xmath70 and @xmath71 are linearly dependent .
this corresponds to exactly one zero - eigenvalue of the jacobian , whereas the other six eigenvalues are stable under the conditions given above .
according to stable manifold theory , for every chosen value of @xmath71 , there exists a stable manifold on which the trajectory converges asymptotically to equilibrium @xmath54 .
( in the following , we have set @xmath66 . ) in this second equilibrium state , the whole swarm is frozen to a static configuration , which satisfies @xmath72 const .
relaxation to a jelly - like state ( @xmath54 ) , according to equations ( [ eins])-([zwei ] ) .
simulation parameters are @xmath60 , @xmath73 , @xmath62 , @xmath63 , @xmath64 and @xmath65 ( @xmath66 ) . ]
one example of such a frozen state is shown in figure [ fig : two ] .
if the swarm is subject to noise influences , the individual particles will oscillate around their fixed positions .
+ + _ dynamic _ states of the swarm are realized when reaching equilibrium @xmath56 , since the sum of all kinetic energies then has a non - vanishing value .
@xmath56 is found to be stable under the conditions @xmath74 , @xmath75 and @xmath76 again , one zero eigenvalue appears and the ruth - hurwitz - theorem was applied to determine the stability of the system .
note that in @xmath56 , the two equations for @xmath70 and @xmath77 define two planes in the space of variables @xmath78 , whose intersection gives a line , i.e. a manifold of equilibria with dimension one , which matches the overall number of zero eigenvalues .
hence the theory of stable manifolds again ensures that we have converging trajectories to @xmath56 under the conditions given above .
if these inequalities are satisfied by a specific parameter choice , the swarm reaches equilibrium @xmath56 and appears to simulate the amorphous behavior of insect swarms .
state of amorphous swarming ( @xmath56 ) , according to equations ( [ eins])-([zwei ] ) .
simulation parameters are @xmath79 , @xmath80 , @xmath62 , @xmath81 , @xmath82 and @xmath83 ( @xmath66 ) . ] in figure [ fig : three ] , the equilibrating process is explicitly shown .
+ + numerical simulation of our original equations ( [ ui])-([oi ] ) gives the full @xmath84-particle time evolution .
video examples for the corresponding motion of swarm particles are available online @xcite for all three distinct cases @xmath53 ( collapse ) , @xmath54 ( jelly - like - state ) and @xmath56 ( amorphous swarming ) .
simulations of the stochastic system ( [ ui])-([oi ] ) in two dimensions are shown , which - apart of small noise perturbations - agree nicely with our simulations of the deterministic equations ( [ eins])-([zwei ] ) .
we have formulated swarm dynamics based on canonical active brownian motion .
we assumed the existence of a single global internal energy of the swarm and postulated its coupling to the full hamiltonian of the system .
we studied the case where all particles were harmonically attracted to their center of mass and were all moving in a landscape of harmonic shape .
the system was simulated numerically and analytic understanding was obtained by studying stability properties of equilibria .
we found three distinct forms of the large time behavior of the swarm , showing collapse to a point , freezing to a static configuration and amorphous swarming behavior , respectively .
+ + an interesting application of our present investigation appears to be the study of particle swarms in given ( genuinely complicated ) potentials and the development of a new algorithm for particle swarm optimization theory .
furthermore we mention the modeling of crystals and applications in solid state physics .
+ + * acknowledgments * : we thank josef hofbauer for his valuable advice on dynamical systems .
in addition , we are grateful for financial support within the agreement on cooperation between the universities of vienna and zagreb . 99 f. schweitzer , w. ebeling , and b. tilch : _ complex motion of brownian particles with energy depots _ , phys
80 * , 5044 ( 1998 ) .
f. schweitzer : _ brownian agents and active particles _ , springer , berlin 2003 .
a. glck , h. hffel , s. iliji : _ canonical active brownian motion _ , phys .
e * 79 * , 021120 ( 2009 ) .
f. schweitzer , w. ebeling , b. tilch : _ statistical mechanics of canonical - dissipative systems and applications to swarm dynamics _ ,
e * 64 * , 021110 ( 2001 ) .
dejesus e. x. , kaufman ch . :
_ routh - hurwitz criterion in the examination of eigenvalues of a system of nonlinear ordinary differential equations _ ,
a * 35 * ( 12 ) , 5288 - 5290 ( 1987 ) . | we present a swarm model of brownian particles with harmonic interactions , where the individuals undergo canonical active brownian motion , i.e. each brownian particle can convert internal energy to mechanical energy of motion .
we assume the existence of a single global internal energy of the system .
numerical simulations show amorphous swarming behavior as well as static configurations .
analytic understanding of the system is provided by studying stability properties of equilibria . |
the workload management task ( work package 1 , or wp1 ) @xcite of the eu datagrid project @xcite ( also known , and referred to in the following text , as edg ) is mandated to define and implement a suitable architecture for distributed scheduling and resource management in the grid environment . during the first year and a half of the project ( 2001 - 2002 ) , and following a technology evaluation process , edg wp1 defined , implemented and deployed a set of services that integrate existing components , mostly from the condor @xcite and globus @xcite projects .
this was described in more detail at chep 2001 @xcite . in a nutshell , the core job submission component of condorg ( @xcite ) ,
talking to computing resources ( known in datagrid as computing elements , or ces ) via the globus gram protocol , is fundamentally complemented by : * a job requirement matchmaking engine ( called the _ resource broker _ , or rb ) , matching job requests to computing resource status coming from the information system and resolving data requirements against the replicated file management services provided by edg wp2 . *
a job logging and book - keeping service ( lb ) , where a job state machine is kept current based on events generated during the job lifetime , and the job status is made available to the submitting user .
the lb events are generated with some redundancy to cover various cases of loss .
* a stable user api ( command line , c++ and java ) for access to the system .
job descriptions are expressed throughout the system using the condor classified ad language , where appropriate conventions were established to express requirement and ranking conditions on computing and storage element info , and to express data requirements .
more details on the structure and evolution of these services and the necessary integration scaffolding can be found in various edg public deliverable documents .
this paper focuses on how the experience of the first year of operation of the wp1 services on the edg testbed was interpreted , digested , and how a few design _ principles _ were learned ( possibly the hard way ) from the design and implementation shortcomings of the first release of wp1 software .
these principles were applied to design and implement the second major release of wp1 software , that is described in another chep 2003 paper ( @xcite ) . to illustrate the logical path that leads to at least some of these principles , we start by exploring the available techniques to model the behaviour and throughput of the integrated workload management system , and identify two factors that significantly complicate the system analysis .
the workload management system provided by edg - wp1 is designed to rely as much as possible on existing technology .
while this has the obvious advantages of limiting effort duplication and facilitating the compatibility among different projects , it also significantly complicates troubleshooting across the various layers of software supplied by different providers , and in general the understanding of the integrated system . also , where negotiations with external software providers could nt reach an agreement within the edg deadlines , some of the interfaces and communication paths in the system had to be adapted to fit the existing external software incarnations . to get a useful high - level picture of the integrated workload management system , beyond all these practical constraints , we can model it as a queuing system , where job requests traverse a network of queues , and the service stations " connected to each queue represent one of the various processing steps in the job life - cycle .
a few of these steps are exemplified in figure [ fig - netqueue ] .
establishing the scale factors for each service in the wp1 system ( e.g. : how many users can a single matchmaking / job submission station serve , how many requests per unit time can a top - level access point to the information system serve , what is the sustained job throughput that can be achieved through the workload management chain , etc . ) is one of the fundamental premises for the correct design of the system .
one could expect to obtain this knowledge either by applying queuing theory to this network model ( this requires obtaining a formal representation of all the components , their service time profiles and their interconnections ) or by measuring the service times and by identifying where long queues are likely to build up when a realistic " request load is injected in the system .
this information could in principle also be used to identify the areas of the system where improvement is needed ( sometimes collectively called _ bottlenecks _ ) .
experience with the wp1 software integration showed that both of these approaches are impractical for either dimensioning the system or ( possibly even more important ) for identifying the trouble areas that affect the system throughput .
we identified two non - linear factors that definitely work against the predictive power of queuing theory in this case , and require extra care even to apply straightforward reasoning when bottlenecks are to be identified to improve system throughput .
these are the consequence of common programming practice ( and are therefore easy to be found in the software components that we build or are integrating ) and are described in the following section .
one of the most common ( and most frustrating , both to developers and to end users ) experiences in troubleshooting the wp1 workload management system on the edg testbed has been the fact that often , perceived _
improvements _ to the system ( sometimes even simple bug fixes ) result in a _ decrease _ in the system stability , or reliability ( fraction of requests that complete successfully ) .
the cause is often closely related to the known fact that removing a bottleneck , in any flow system , can cause an overflow downstream , possibly close to the _ next _ bottleneck .
the complicating factor is that there are at least two characteristics that could ( and possibly still can ) be found in many elements of our integrated workload management queuing network , that can cause problems to appear even very far from the area of the network where an _ improvement _ is being attempted : * _ queues of job requested can form where they can impact on the system load . _ + different techniques can be chosen or needed to pass job requests around . sometimes a socket connection is needed , sometimes sequential request processing ( one request at a time in the system ) is required for some reason , and multiple processes / threads may be used to handle individual requests . having a number of tasks ( processes / threads ) wait for a socket queue or a sequential processing slot
is one way to queue " requests that definitely generates much extra work for the process scheduler , and can cause any other process served by the same scheduler to be allocated less and less time .
queues that are unnecessarily scanned while waiting for some other condition to allow the processing of their element can also impact on the system load , especially if the queue elements are associated to significant amounts of allocated dynamic memory . *
_ some system components can enforce hard timeouts and cause anomalies in the job flow . _
+ when handling the access ( typically via socket connections ) to various distributed services , provisions typically need to be made to handle all possible failure modes .
reasonably " long timeouts are sometimes chosen to handle failures that are perceived to be very unlikely by developers ( failure to establish communication to a local service , for instance ) .
this kind of failures , however , can easily materialise when the system resources are exhausted under a stress test or load peak .
figure [ fig - fmode ] illustrates how these two effects can conspire to frustrate a genuine effort to remove what seems the limiting bottleneck in the system ( the example in the figure does nor refer to any real case or component ) : removing the bottleneck ( 1 ) causes a request queue to build up at the next station ( 2 ) , and this interferes via the system load to cause hard timeouts and job failures elsewhere ( 3 ) .
this example is used to rationalise some of the unexpected reactions that , in many cases , were found while working on the wp1 integrated system .
the experience on practical troubleshooting cases similar to this one , while bringing an understanding of the difficulties inherent in building distributed systems , also drove us to formulate some of the principles that are presented in the next section .
the attempts at getting a deeper understanding of the edg - wp1 workload management system and their failures led us to formulate a few design principles and to apply them to the second major software release . here
are the principles that descend from the paradigm example described in section [ sec - fmode ] : 1 .
* queues of various kinds of requests for processing should be allowed to form where they have a minimal and understood impact on system resources . *
+ queues that get ` filled ' in the form of multiple threads or processes , or that allocate significant amounts of system memory should be avoided , as they not only adversely impact system performance , but also generate inter - dependencies and complicate troubleshooting .
* limits should always be placed on dynamically allocated objects , threads and/or subprocesses . *
+ this is a consequence of the previous point : every dynamic resource that gets allocated should have a tunable system - wide limit that gets enforced .
* special care needs to be taken around the pipeline areas where serial handling of requests is needed .
* + the impact of any contention for system resources becomes more evident near areas of the queuing system that require the acquisition of system - wide locks .
so far we concentrated on a specific attempt at modeling and understanding the workload management system that led to an increased attention to the usage of shared resources .
there were other specific practical issues that emerged during the deployment and troubleshooting of the system and that led to the awareness of some fundamental design or implementation mistake that was made .
here is a short list , where the fundamental principle that should correct the fundamental mistake that was made is listed : 1 .
* communication among services should always be reliable : * * always applying double - commit and rollback for network communications . * going through the filesystem for local communications .
+ in general , forms of communication that do nt allow for data or messages to be lost in a broken pipe lead to easier recovery from system or process crashes . where network communication is necessary
, database - like techniques have to be used .
every process , object or entity related to the job lifecycle should have another process , object or entity in charge of its well - being . *
+ automatic fault recovery can only happen if every entity is held accountable and accounted for .
* information repositories should be minimized ( with a clear identification of authoritative information ) . *
+ many of the software components that were integrated in the edg - wp1 solution are stateful and include local repositories for request information , in the form of local queues , state files , database back - ends .
only one site with authoritative information about requests has to be identified and kept .
monolithic , long - lived processes should be avoided . * + dynamic memory programming , using languages and techniques that require explicit release of dynamically allocated objects , can lead to leaks of memory , descriptors and other resources .
experimental , r&d code can take time to leak - proof , so it should possibly not be linked to system components that are long - lived , as it can accelerate system resource starvation . short - lived , easy - to - recover components are a clean and very practical workaround in this case
more thought should be devoted to efficiently and correctly recovering a service rather than to starting and running it . *
+ this is again a consequence of the previous point : the capability to quickly recover from failures or interruption helps in assuring that system components ` can ' be short - lived , either by design or by accident .
edg - wp1 has been distributing jobs over the edg testbed in a continuous fashion for one and a half years now , with a software solution where existing grid technology was integrated wherever possible .
the experience of understanding the direct and indirect interplay of the service components could not be reduced to a simple _ scalability _ evaluation .
this because understanding and removing _ bottlenecks _ is significantly complicated by non - linear and non - continuous effects in the system . in this process
, few principles that apply to the very complex practice of distributed systems operations were learned the hard way ( i.e. not by just reading some good book on the subject ) .
edg - wp1 tried to incorporate these principles in its second major software release that will shortly face deployment in the edg testbed .
j. frey , t. tannenbaum , i. foster , m. livny , s. tuecke , `` condor - g : a computation management agent for multi - institutional grids '' , _ proceedings of the tenth ieee symposium on high performance distributed computing ( hpdc10 ) _ , 2001 datagrid wp1 members ( c. anglano _ et al .
_ ) , * integrating grid tools to build a computing resource broker : activities of datagrid wp1 * " presented at the chep 2001 conference , beijing ( p. 708
in the proceedings ) | application users have now been experiencing for about a year with the standardized resource brokering services provided by the workload management package of the eu datagrid project ( wp1 ) .
understanding , shaping and pushing the limits of the system has provided valuable feedback on both its design and implementation .
a digest of the lessons , and better practices " , that were learned , and that were applied towards the second major release of the software , is given . |
we often see published results in the form @xmath0 where @xmath1 and @xmath2 are _ usually _ positive . and
@xmath2 with all combinations of signs , see public online tables of deep inelastic scattering results.@xcite i want to make clear since the very beginning that it is not my intention to blame experimental or theoretical teams which have reported in the past asymmetric uncertainty , because we are all victims of a bad tradition in data analysis .
at least , when asymmetric uncertainties have been given , there is some chance to correct the result , as described in sec .
[ sec : thumb ] . since some asymmetric contributions to the global uncertainties almost unavoidably happen in complex experiments , i am more worried of collaborations that never arrive to final asymmetric uncertainties , because i must imagine they have symmetrised somehow the result but , i am afraid , without applying the proper shifts to the ` best value ' to take into account asymmetric contributions , as it will be discussed in the present paper . ] as firstly pointed out in ref .
@xcite and discussed in a simpler but more comprehensive way in ref .
@xcite , this practice is far from being acceptable and , indeed , could bias the believed value of important physics quantities .
the purpose of the present paper is , summarizing and somewhat completing the work done in the above references , to remind where asymmetric uncertainty stem from and to show why , as they are usually treated , they bias the value of physical quantities , either in the published result itself or in subsequent analyses .
once the problems are spotted , the remedy is straightforward , at least within the bayesian framework ( see e.g. @xcite , or @xcite and @xcite for recent reviews ) .
in fact the bayesian approach is conceptually based on the intuitive idea of probability , and formally grounded on the basic rules of probability ( what are usually known as the probability ` axioms ' and the ` conditional probability definition ' ) plus logic . within this framework
many methods of ` conventional ' statistics are reobtained , as approximations of general solutions , under well stated conditions of validity . instead , in the conventional , frequentistic approach _
ad hoc _ formulae , prescriptions and un - needed principles are used , often without understanding what is behind these methods before a ` principle ' there is nothing !
the proposed bayesian solutions to cure the troubles produced by the usual treatment of asymmetric uncertainties is to step up from approximated methods to the more general ones ( see e.g. ref .
@xcite , in particular the top down approximation diagram of fig .
2.2 ) . in this paper
we shall see , for example , how @xmath3 and minus log - likelihood fit ` rules ' can be derived from the bayesian inference formulae as approximated methods and what to do when the underlying conditions do not hold .
we shall encounter a similar situation regarding standard formulae to propagate uncertainty .
some of the issues addressed here and in refs . @xcite and @xcite have been recently brought to our attention by roger barlow @xcite , who proposes frequentistic ways out .
michael schmelling had also addressed questions related to ` asymmetric errors ' , particularly related to the issue of weighted averages @xcite .
the reader is encouraged to read also these references to form his / her idea about the spotted problems and the proposed solutions . in sec .
[ sec : propagation ] the issue of propagation of uncertainty is briefly reviewed at an elementary level ( just focusing on the sum of uncertain independent variables i.e. no correlations considered ) though taking into account asymmetry in probability density functions ( p.d.f . ) of the _ input _ quantities . in this way we understand what ` might have been done ' ( we are rarely in the positions to exactly know `` what has been done '' ) by the authors who publish asymmetric results and
what is the danger of improper use of such a published ` best value ' _ as is _ in subsequent analyses .
then , sec .
[ sec : sources ] we shall see in where asymmetric uncertainties stem from and what to do in order to overcome their potential troubles . this will be done in an exact way and , whenever is possible , in an approximated way . some rules of thumb to roughly recover sensible probabilistic quantities ( expected value and standard deviation ) from results published with asymmetric uncertainties will be given in sec .
[ sec : thumb ] .
finally , some conclusions will be drawn .
determining the value of a physics quantity is seldom an end in itself . in most cases
the result is used , together with other experimental and theoretical quantities , to calculate the value of other quantities of interest . as it is well understood , uncertainty on the value of each ingredient
is propagated into uncertainty on the final result .
if uncertainty is quantified by probability , as it is commonly done explicitly or implicitly in physics , the propagation of uncertainty is performed using rules based on probability theory .
if we indicate by @xmath4 the set ( ` vector ' ) of input quantities and by @xmath5 the final quantity , given by the function @xmath6 of the input quantities , the most general propagation formula ( see e.g. @xcite ) is given by ( we stick to continuous variables ) : @xmath7\cdot f(\mvec x)\ , \mbox{d}\mvec x\ , , \label{eq : prop_general}\ ] ] where @xmath8 is the p.d.f .
of @xmath5 , @xmath9 stands for the joint p.d.f . of @xmath4 and @xmath10
is the dirac delta ( note the use of capital letters to name variables and small letters to indicate the values that variables may assume ) .
the exact evaluation of eq .
( [ eq : prop_general ] ) is often challenging , but , as discussed in ref .
@xcite , this formula has a nice simple interpretation that makes its monte carlo implementation conceptually easy .
as it is also well known , often there is no need to go through the analytic , numerical or monte carlo evaluation of eq.([eq : prop_general ] ) , since linearization of @xmath11 around the expected value of @xmath4 ( e[@xmath4 ] ) makes the calculation of expected value and variance of @xmath5 very easy , using the well known standard propagation formulae , that for uncorrelated input quantities are @xmath13 & \approx & y(\mbox{e}[\mvec x ] ) \label{eq : prop_approx_e } \\
\sigma^2(y ) & \approx & \sum_i \left(\left.\frac{\partial y}{\partial x_i } \right|_{\mbox{e}[\mvec x]}\right)^2\ , \sigma^2(x_i)\ , .
\label{eq : prop_approx_sigma}\end{aligned}\ ] ] as far as the shape of @xmath8 , a gaussian one is usually assumed , as a result of the central limit theorem .
holding this assumptions , @xmath14 $ ] and @xmath15 is all what we need .
@xmath14 $ ] gives the ` best value ' , and probability intervals , upper / lower limits and so on can be easily calculated .
in particular , within the gaussian approximation , the most believable value ( _ mode _ ) , the barycenter of the p.d.f .
( _ expected value _ ) and the value that separates two adjacent 50% probability intervals ( _ median _ ) coincide . if @xmath8 is asymmetric this is not any longer true and one needs then to clarify what ` best value ' means , which could be one of the above three _ position parameters _ , or something else ( in the bayesian approach ` best value ' stands for expected value , unless differently specified ) . anyhow , gaussian approximation is not the main issue here and , in most real applications , characterized by several contributions to the combined uncertainty about @xmath5 , this approximation is a reasonable one , even when some of the input quantities individually contribute asymmetrically .
my concerns in this paper are more related to the evaluation of @xmath14 $ ] and @xmath15 when 1 . instead of eqs .
( [ eq : prop_approx_e])([eq : prop_approx_sigma ] ) , _ ad hoc _ propagation prescriptions are used in presence of asymmetric uncertainties ; 2 .
linearization implicit in eqs .
( [ eq : prop_approx_e])([eq : prop_approx_sigma ] ) is not a good approximation .
let us start with the first point , considering , as an easy academic example , input quantities described by the asymmetric triangular distribution shown in the left plot of fig .
[ fig:2triang ] . [ cols="^,^,^,^ " , ] a non parabolic ,
asymmetric @xmath3 produces an asymmetric @xmath16 ( see fig . [
fig : asymmetric_chi2 ] ) , the mode of which corresponds , indeed , to what obtained minimizing @xmath3 , but expected value and standard deviation differ from what is obtained by the ` standard rule ' . in particular , expected value and variance must be evaluated from their definitions : @xmath17 & = & \int\!\theta\,f(\theta\,|\,\mbox{data})\,\mbox{d}\theta \label{eq : theta_e } \\
\sigma^2_\theta & = & \int\!(\theta-\mbox{e}[\theta])^2 \,f(\theta\,|\,\mbox{data})\,\mbox{d}\theta\ , .
\label{eq : theta_sigma}\end{aligned}\ ] ] other examples of asymmetric @xmath3 curves , including the case with more than one minimum , are shown in chapter 12 of ref .
@xcite , and compared with the results coming from frequentist prescriptions ( but , indeed , there is not a general accepted rule to get frequentistic results
whatever they mean when the @xmath3 shape gets complicated ) . unfortunately , it is not easy to translate numbers obtained by _ ad hoc _ rules into probabilistic results , because the dependence on the actual shape of the @xmath3 or @xmath18 curve can be not trivial .
anyhow , some _ rules of thumb _ can be given in next - to - simple situations where the @xmath3 or @xmath18 has only one minimum and the @xmath3 or @xmath18 curve looks like a ` skewed parabola ' , like in fig .
[ fig : asymmetric_chi2 ] : * the 68% ` confidence interval ' obtained by the @xmath19 , or @xmath20 rule still provides a 68% probability interval for @xmath21 .
* the standard deviation obtained using eq .
( [ eq : theta_sigma ] ) is approximately equal to the average between the @xmath1 and @xmath2 values obtained by the @xmath19 , or @xmath20 rule : @xmath22 * the expected value is equal to the mode ( @xmath23 , coinciding with the maximum likelihood or minimum @xmath3 value ) plus a _ shift _ : @xmath24 \approx \theta_m + { \cal o}(\delta_+ - \delta_- ) \ , .
\label{eq : shift_chi2}\ ] ] [ this latter rule is particularly rough because @xmath25 $ ] is more sensitive than @xmath26 on the exact shape of @xmath3 or @xmath18 curve . equation ( [ eq : shift_chi2 ] ) has to be taken only to get an idea of the order of magnitude of the effect .
for example , in the case depicted in fig [ fig : asymmetric_chi2 ] the shift is 80% of @xmath27 . ] the remarks about misuse of @xmath19 and @xmath20 rules can be extended to cases where several parameters are involved .
i do not want to go into details ( in the bayesian approach there is nothing deeper than studying @xmath28 or @xmath29 in function of several parameters .. ] ) , but i just want to get the reader worried about the meaning of contour plots of the kind shown in fig .
[ fig : spots ] .
another source of asymmetric uncertainties is nonlinear dependence of the output quantity @xmath5 on some of the input @xmath4 in a region a few standard deviations around @xmath30 .
this problem has been studied with great detail in ref .
@xcite , also taking into account correlations on input and output quantities , and somewhat summarized in ref . @xcite .
let us recall here only the most relevant outcomes , in the simplest case of only one output quantity @xmath5 and neglecting correlations .
+ & + & + figure [ fig : quad ] shows a non linear dependence between @xmath31 and @xmath5 and how a gaussian distribution has been distorted by the transformation [ @xmath8 has been evaluated analytically using eq.([eq : prop_general ] ) ] . as a result of the nonlinear transformation , mode , mean ,
median and standard deviation are transformed in non trivial ways ( in the example of fig .
[ fig : quad ] mode moves left and expected value right ) . in the general case
the complete calculations should be performed , either analytically , or numerically or by monte carlo .
fortunately , as it has been shown in ref .
@xcite , second order expansion is often enough to take into account small deviations from linearity .
the resulting formulae are still compact and depend on location and shape parameters of the initial distributions .
second order propagation formulae depend on first and second derivatives . in practical cases ( especially as far as the contribution from systematic effects
are concerned ) the derivatives are obtained numerically as @xmath32 } & \approx & \frac{1}{2 } \left(\frac{\delta _ + } { \sigma(x)}+ \frac{\delta _ -}{\sigma(x)}\right ) = \frac{\delta _
+ + \delta _ -}{2\,\sigma(x)}\ , , \\ \left.\frac{\partial^2 y}{\partial x^2}\right|_{\mbox{\small e}[x ] } & \approx & \frac{1}{\sigma(x)}\ , \left(\frac{\delta _ + } { \sigma(x)}-\frac{\delta _ -}{\sigma(x)}\right ) = \frac{\delta _ + -\delta _ -}{\sigma^2(x)}\,,\end{aligned}\ ] ] where @xmath2 and @xmath1 now stand for the left and right deviations of @xmath5 when the _ input variable @xmath31 varies by one standard deviation _ around @xmath33
second order propagation formulae are conveniently given in ref .
@xcite in terms of the @xmath34 deviations and @xmath35 are given by @xmath36 } \ , \sigma^2(x ) \\
\overline{\delta } & = & \left.\frac{\partial y } { \partial x}\right|_{\mbox{\small e}[x ] } \ , \sigma(x)\,.\end{aligned}\ ] ] ] . for @xmath5 that depends only on a single input @xmath31
we get : @xmath37 ) + \delta\ , , \label{eq : ey_nonlinear } \\
\sigma^2(y ) & \approx & \overline{\delta}^2 + 2\,\overline{\delta}\cdot\delta\cdot s(x)+ \delta^2\cdot\left[{\cal k}(x)-1\right]\ , , \label{eq : sig_nonlinear}\end{aligned}\ ] ] where @xmath10 is the semi - difference of the two deviations and @xmath35 is their average : @xmath38 while @xmath39 and @xmath40 stand for skewness and kurtosis of the input variable
. we should not forget that the input quantities could have non trivial shapes . since skewness and kurtosis are related to 3rd and 4th moment of the distribution , eq .
( [ eq : sig_nonlinear ] ) makes use up to the 4th moment and is definitely better that the usual propagation formula , that uses only second moments . in ref .
@xcite approximated formulae are given also for skewness and kurtosis of the output variable , from which it is possible to reconstruct @xmath8 taking into account up to 4-th order moment of the distribution . ] for many input quantities we have @xmath37 ) + \sum_i\delta_i\ , , \label{eq : ey_nonlinear_many } \\
\sigma^2(y ) & \approx & \sum_i\sigma^2_{x_i}(y)\ , , \label{eq : sig_nonlinear_many}\end{aligned}\ ] ] where @xmath41 stands for each individual contribution to eq .
( [ eq : sig_nonlinear ] ) .
the expression of the variance gets simplified when all input quantities are gaussian ( a gaussian has skewness equal 0 and kurtosis equal 3 ) : @xmath42 and , as long as @xmath43 are much smaller that @xmath44 , we get the convenient _ approximated formulae _
@xmath45 ) + \sum_i \delta_i\ , , \label{eq : nl_simple_e } \\
\sigma^2(y ) & \approx & \sum_i \overline{\delta}^2_i\,\ , \label{eq : nl_simple}\end{aligned}\ ] ] valid also for other symmetric input p.d.f.s ( the kurtosis is about 2 to 3 in typical distribution and its exact value is irrelevant if the condition @xmath46 holds )
. the resulting practical rules ( [ eq : nl_simple_e])([eq : nl_simple ] ) are quite simple : * the _ expected value _ of @xmath5 _ is shifted _ by the sum of the individual shifts , each given by half of the semi - difference of the deviations @xmath34 ; * each input quantity contributes ( in quadrature ) to the combined standard uncertainty with a term which is approximately the average between the deviations @xmath34 .
moreover , if there are many contributions to the uncertainty , the final uncertainty will be symmetric and approximately gaussian , thanks to the central limit theorem . finally , and this is often the case that we see in publications , asymmetric uncertainty results from systematic effects .
the bayesian approach offers a natural and clear way to treat systematics and
i smile at the many attempts of ` squaring the circle ' using frequentistic prescriptions
simply because probabilistic concepts are consistently applied to all _ influence quantities _ that can have an effect on the quantity of interest and whose value is not precisely known .
therefore we can treat them using probabilistic methods .
this was also recognized by metrologic organizations@xcite .
indeed , there is no need to treat systematic effects in a special way .
they are treated as any of the many input quantities @xmath4 discussed in sec .
[ ss : nonlinear ] , and , in fact , their asymmetric contributions come frequently from their nonlinear influence on the quantity of interest . the only word of caution , on which i would like to insist , is to use expected value and standard deviation for each systematic effect .
in fact , sometimes the uncertainty about the value of the influence quantities that contribute to systematics is intrinsically asymmetric .
i also would like to comment shortly on results where either of the @xmath34 is negative , for example @xmath47 ( see e.g. ref .
@xcite to have an idea of the variety of signs of @xmath34 ) .
this means that that the we are in proximity of a minimum ( or a maximum if @xmath1 were negative ) of the function @xmath48 .
it can be shown @xcite that eqs .
( [ eq : ey_nonlinear])-([eq : sig_nonlinear ] ) hold for this case too . by @xmath49 around its expected value @xmath50 $ ] the final quantity @xmath5 only moves in one side of @xmath51)$ ] . ] for further details about meaning and treatment of uncertainties due systematics and their relations to iso _ type b _
uncertainties@xcite , see refs .
@xcite and @xcite .
having understood what one should have done to obtain expected value and standard deviation in the situations in which people are used to report asymmetric uncertainties , we might attempt to recover those quantities from the published result .
it is possible to do it exactly only if we know the detailed contributions to the uncertainty , namely the @xmath3 or log - likelihood functions of the so called ` statistical terms ' and the pairs @xmath52 , together to the probabilistic model , for each ` systematic term ' . however , these pieces of information are usually unavailable .
but we can still make some _ guesses _ , based on some rough assumptions , lacking other information : * asymmetric uncertainties in the ` statistical part ' are due to asymmetric @xmath3 or log - likelihood : @xmath53 apply corrections given by eqs .
( [ eq : sigma_delta_chi2])([eq : shift_chi2 ] ) ; * asymmetric uncertainties in the ` systematic part ' comes from nonlinear propagation : @xmath53 apply corrections given by eqs .
( [ eq : nl_simple_e])([eq : nl_simple ] ) . as a numerical example , imagine we read the following result ( in arbitrary units ) : @xmath54 ( that somebody would summary as @xmath55 ! ) .
the only certainty we have , seeing two asymmetric uncertainties with the same sign of skewness , is that _ the result is definitively biased_. let us try to make our estimate of the bias and calculate the corrected result ( that , not withstanding all uncertainties about uncertainties , will be closer to the ` truth ' than the published one ) : 1 .
the first contribution gives roughly [ see .
( [ eq : sigma_delta_chi2])([eq : shift_chi2 ] ) ] : @xmath56 2 . for the second contribution we have
[ see . eqs .
( [ eq : delta_m])([eq : delta_m ] ) , ( [ eq : nl_simple_e])([eq : nl_simple ] ) ] : @xmath57 _ our _ guessed best result would then become notation . in this example
we would have @xmath58 .
personally , i do not think this is a very important issue as long as we know what the quantity @xmath59 means .
anyhow , i understand the iso rational , and perhaps the proposed notation could help to make a break with the ` confidence intervals ' . ]
@xmath60 ( the exceeding number of digits in the intermediate steps are just to make numerical comparison with the correct result that will be given in a while . ) if we had the chance to learn that the result of eq .
( [ eq : esempio_bad_y ] ) was due to the asymmetric @xmath3 fit of fig .
[ fig : asymmetric_chi2 ] plus two systematic corrections , each described by the triangular distribution of fig .
[ fig:2triang ] , then we could calculate expectation and variance exactly : @xmath61 i.e. @xmath62 , quite different from eq .
( [ eq : esempio_bad_y ] ) and close to the result corrected by rule of thumb formulae
. indeed , knowing exactly the ingredients , we can evaluate @xmath8 from eq.([eq : prop_general ] ) as @xmath63 although by monte carlo .
the result is given in fig .
[ fig : skewed+2tr ] , from which we can evaluate a mean value of 4.54 and a standard deviation of 1.65 in perfect agreement with the figures given in eqs .
( [ eq : exact_corr_e])([eq : exact_corr_sigma ] ) . of fig .
[ fig : asymmetric_chi2 ] is the rounded value of 1.54 .
replacing 1.5 by 1.54 in eq .
( [ eq : exact_corr_sigma ] ) , we get exactly the monte carlo value of 1.65 . ] as we can see from the figure , also those who like to think at best value in term of most probable value have to realize once more that _
the most probable value of a sum is not necessarily equal to the sum of most probable values of the addends _ ( and analogous statements for all combinations of uncertainties ) . in the distribution of fig .
[ fig : skewed+2tr ] , the mode of the distribution is around 5 . [ note that expected value and variance are equal to those given by eqs .
( [ eq : exact_corr_e])([eq : exact_corr_sigma ] , since in the case of a linear combination they can be obtained exactly . ]
other statistical quantities that can be extracted by the distribution are the median , equal to 4.67 , and some quantiles ( values at which the cumulative distribution reaches a given percent of the maximum the median being the 50% quantile ) .
interesting quantiles are the 15.85% , 25% , 75% and 84.15% , for which the monte carlo gives the following values of @xmath5 : 2.88 , 3.49 , 5.72 and 6.18 . from these values
we can calculate the _ central _ 50% and 68.3% intervals , ] is almost unavoidable ( i have known physicists convinced and who even taught it ! that the standard deviation only ` makes sense for the gaussian ' and that it was defined via the ` 68% rule ' ) .
for this reason , recently i have started to appreciate thinking in terms of 50% probability intervals , also because they force people to reason in terms of better perceived fifty - to - fifty bets .
i find these kind of bets very enlighting to show why practically all standard ways ( including bayesian ones ! ) fail to report upper / lower _ confidence limits _ in _ frontier case situations
_ characterized by _
open likelihoods _ ( see chapter 12 in ref.@xcite ) .
i like to ask `` please use your method and give me a 50% c.l .
upper / lower limit '' , and then , when i have got it , `` are you really 50% confident that the value is below that limit and 50% confident that it is above it
? would you equally bet on either side of that limit ? '' . and the supporters of ` objective ' methods are immediately at loss .
( at least those who use bayesian formulae realize that there must be some problem with the choice of priors . ) ] which are @xmath64 $ ] and @xmath65 $ ] , respectively . again , the information provided by eq .
( [ eq : esempio_bad_y ] ) is far from any reasonable way to provide the uncertainty about @xmath5 , given the information on each component . besides the lucky case of this numerical example ( which was not constructed on purpose , but just recycling some material from ref .
@xcite ) , it seems reasonable that even results roughly corrected by rule of thumb formulae are already better than those published directly with asymmetric result .
was @xmath66 , without further information , we could still try to apply some shift to the result , obtaining @xmath67 or @xmath68 depending on some guesses about the source of the asymmetry . in any case , either results are better than @xmath66 ! ] but the accurate analysis can only be done by the authors who know the details of the individual contribution to the uncertainty .
asymmetric uncertainties do exist and there is _ no way to remove them artificially_. if they are not properly treated , i.e. using prescriptions that do not have a theoretical ground but are more or less rooted in the physics community , the published result is biased . instead ,
if they are properly treated using probability theory , in most cases of interest the final result is practically symmetric and approximately gaussian , with expected value and standard deviations which take into account the several shifts due to individual asymmetric contributions .
note that some of the simplified methods to make statistical analyses had a _
raison dtre _ many years ago , when the computation was a serious limitation .
now it is not any longer a problem to evaluate , analytically or numerically , integrals of the kind of those appearing e.g. in eqs.([eq : prop_general ] ) , ( [ eq : theta_e ] ) and ( [ eq : theta_sigma ] ) .
in the case the final uncertainty remains asymmetric , the authors should provide detailed information about the ` shape of the uncertainty ' , giving also most probable value , probability intervals , and so on .
but the best estimate of the _ expected value and standard deviation should be always given _
( see also the _ iso guide _
@xcite ) . to conclude
, i would like to leave the final word to my preferred quotation with whom i like to end seminars and courses on probability theory applied to the evaluation and the expression of uncertainty in measurements : _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ `` although this guide provides a framework for assessing uncertainty , it can not substitute for critical thinking , intellectual honesty , and professional skill .
the evaluation of uncertainty is neither a routine task nor a purely mathematical one ; it depends on detailed knowledge of the nature of the measurand and of the measurement .
the quality and utility of the uncertainty quoted for the result of a measurement therefore ultimately depend on the understanding , critical analysis , and integrity of those who contribute to the assignment of its value.''_@xcite _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ h1 collaboration , www-h1.desy.de/psfiles/figures/d97-042.tab_corre + lations.txt ; + zeus collaboration , www-zeus.desy.de/physics/sfew/public/ + sfew_results / published / nc-9900/table-2.dat .
g. dagostini and m. raso , cern - ep/2000 - 026 , february 2000 [ hep - ex/0002056 ] .
g. dagostini , _ bayesian reasoning in data analysis : a critical introduction _ , world scientific , 2003 ( book info , including list of contents and hypertexted bibliography can be found at www.roma1.infn.it/@xmath69dagos/wspc/ ) .
g. dagostini , proc .
xviii international workshop on maximum entropy and bayesian methods , garching ( germany ) , july 1998 , v. dose et al .
, kluwer academic publishers , dordrecht , 1999 , 157170 [ physics/9811046 ] . | the issue of asymmetric uncertainties resulting from fits , nonlinear propagation and systematic effects is reviewed .
it is shown that , in all cases , whenever a published result is given with asymmetric uncertainties , the _ value _ of the physical quantity of interest _ is biased _ with respect to what would be obtained using at best all experimental and theoretical information that contribute to evaluate the combined uncertainty .
the probabilistic solution to the problem is provided both in exact and in approximated forms .
# 1 |
mannheim @xcite ( henceforth m6 ) has developed a cosmological model ( mm ) based on conformal invariance .
this model has a number of attractive features , but has not so far been accepted as a viable candidate for the correct theory of gravitation , in large part because it predicts , apparently unambiguously , a negative value for the effective gravitational constant , @xmath0 . from this
follows a thermal history markedly different from the usual one , and consequent difficulties in addressing questions such as primordial nucleosynthesis @xcite . in this paper
we point out a mathematical problem in the formulation of the model , and present a new approach that yields a positive @xmath0 , so that the model closely resembles the conventional one .
the mm involves a scalar field , @xmath1 , in a friedmann - robertson - walker ( frw ) background metric .
it is this field @xmath1 that we will be concentrating on in this paper .
the treatment of both @xmath1 and the gravitational field in m6 is classical .
important subsequent work has been done on the quantization of the fourth - order equations for the gravitational field that result from the conformal lagrangian @xcite .
the scalar field , @xmath1 , however , has always been treated classically , a procedure that we will question in this paper .
there are several things that are required of @xmath1 in a viable model : @xmath2 an equation of motion that is conformally invariant .
an energy - momentum tensor whose 00 component ( hamiltonian ) has an energy spectrum that is bounded below . a constant vacuum expectation value , @xmath3 , derived from the equation of motion .
the @xmath4 in the action generates the effective gravitational constant , @xmath0 . a @xmath5 term in the action that gives the correct sign for the cosmological constant .
let us see whether these criteria are met in the mm . using ( as mannheim does ) a metric signature @xmath6 , and neglecting the coupling to the fermion field , the terms in mannheim s lagrangian involving @xmath1 are ( m6 ( 61 ) ) : @xmath7 this results in an equation of motion ( m6 ( 63 ) ) : @xmath8 this equation , as required ,
is conformally invariant .
@xmath1 is assumed to acquire a non - zero vacuum expectation value , @xmath3 , found by setting the derivative term in ( [ eq : s ] ) equal to zero .
this gives @xmath9 as in @xcite ( 13 ) , with @xmath10 .
further development of the model ( m6 , section 10 ) shows that , in order to meet observational criteria at the present time , @xmath11 and @xmath12 are both negative , so that @xmath4 is real and positive .
the effective gravitational constant , @xmath0 , in the mm turns out to be negative ( m6 ( 224 ) ) : @xmath13 choosing units with @xmath14 .
the stress tensor in the mm model is obtained by taking the variation of the action with respect to the metric , in the usual way ( m6 ( 64 ) ; a similar expression is given in @xcite ) . retaining just the terms involving @xmath1 we get @xmath15 the lagrangian ( [ eq : l ] ) is analogous to the minkowski space lagrangian for a @xmath16 field theory ( @xcite ( 7.2.14 ) , with @xmath17 ) : @xmath18 the @xmath19 term in ( [ eq : l ] ) has the `` right '' sign for an @xmath20 term . but
with @xmath21 , the @xmath22 term has the `` wrong '' sign , and , in a conventional treatment , will lead to a spectrum that is not bounded below .
a theory of this sort has to be treated by methods appropriate to non - hermitean hamiltonians @xcite . at the level of quantum mechanics ,
`` wrong sign '' @xmath16 theory is well developed .
nevertheless , even in minkowski space , constructing a corresponding quantum field theory is a difficult problem , still incompletely understood @xcite .
conformal cosmology will remain flawed until we can make progress in understanding the scalar field .
a conformally invariant theory with a single scalar field has a unique action , m6 ( 61 ) , provided we use the familiar techniques appropriate for hermitean lagrangians .
once we recognize that our theory involves a non - hermitean lagrangian , however , a new approach is suggested , that we introduce in this section .
we begin by working in minkowski space , but retain @xmath23 and @xmath24 in formulae to simplify the transition to a frw space .
a @xmath16 theory with a `` wrong sign '' @xmath16 term is non - hermitean but is nevertheless @xmath25 symmetric , and can be treated by the methods outlined in @xcite .
the distinctive feature of this approach is the use of the @xmath26 norm in place of the usual dirac norm ; for a quantized field , @xmath27 , we write this norm as @xmath28 here @xmath29 and @xmath30 represent the usual parity and time - reversal operations , while @xmath31 represents a special operation designed to ensure the norm is real and positive definite and the theory is unitary .
the @xmath31 operator has to be specifically calculated for each hamiltonian .
our cosmological model is written in terms of classical fields ( expectation values , @xmath32 ) , which we take to be real .
we assume @xmath33 can then be expressed as @xmath34_{\forall \rho : z^{\rho } = -y^{\rho } } \label{eq : cptdef}\end{aligned}\ ] ] ( compare @xcite ( 78 ) ) , so that @xmath35_{\forall \rho : z^{\rho } = -y^{\rho } } s ( x ) \label{eq : normex2}\end{aligned}\ ] ] in an expression of this sort , @xmath36 and @xmath37 describe the field @xmath1 at the same physical point , but use different coordinate systems to refer to that point .
take the complex conjugate of ( [ eq : normex2 ] ) , and let @xmath38 and @xmath39 : @xmath40_{\forall \rho : z^{\rho } = x^{\rho } } \left [ s^ { * } ( u ) \right]_{\forall \rho : u^{\rho } = -y^{\rho } } \nonumber \\ & = & \int { \rmd}^4 x \int { \rmd}^4 y \ , c^ { * } ( x^{\mu } - y^{\mu } ) s^ { * } ( -y ) s(x)\end{aligned}\ ] ] showing that @xmath41 must be real .
we define our action by @xmath42^{\cal cpt } \frac{\partial s ( x)}{\partial x^{\nu } } \right .
\nonumber \\ & & \left .
{ } + \frac{\sigma_m}{2 } m^2 s^{\cal cpt}(x ) s(x ) + \sigma_{\mu } { \mu}^2 \left [ s^{\cal cpt}(x ) s(x ) \right]^2 \right\ } \label{eq : imink } \end{aligned}\ ] ] where @xmath43 , @xmath44 and @xmath45 are simply `` sign factors '' , each of which can be equal to @xmath46 .
we will determine their actual values as we proceed .
the energy - momentum tensor is obtained in the usual way by varying the action with respect to the metric : @xmath47^{\cal cpt } \frac{\partial s ( x)}{\partial x^{\beta } } \right .
\nonumber \\ & & \left .
{ } + \frac{\sigma_m}{2 } m^2 s^{\cal cpt } ( x ) s(x ) + \sigma_{\mu } { \mu}^2 \left [ s^{\cal cpt } ( x )
s(x ) \right]^2 \right\ } \nonumber \\ & & \left .
{ } - ( -g ) ^{1/2 } \frac{\sigma_k}{2 } g^{\mu \alpha } g^{\nu \beta } \left [ \frac{\partial s ( x)}{\partial x^{\alpha } } \right]^{\cal cpt } \frac{\partial s ( x)}{\partial x^{\beta } } \right ) \nonumber \\ & = & g^{\mu \nu } \left\ { \frac{\sigma_k}{2 } g^{\alpha \beta } \left [ \frac{\partial s ( x)}{\partial x^{\alpha } } \right]^{\cal cpt } \frac{\partial s ( x)}{\partial x^{\beta } } \right .
\nonumber \\ & & \left . { }
+ \frac{\sigma_m}{2 } m^2 s^{\cal cpt } ( x ) s(x ) + \sigma_{\mu } { \mu}^2 \left [ s^{\cal cpt } ( x ) s(x ) \right]^2 \right\ } \nonumber \\ & & { } - \sigma_k g^{\mu \alpha } g^{\nu \beta } \left [ \frac{\partial s ( x)}{\partial x^{\alpha } } \right]^{\cal cpt } \frac{\partial s ( x)}{\partial x^{\beta}}\end{aligned}\ ] ] setting @xmath48 , the hamiltonian is given by @xmath49^{\cal cpt } \frac{\partial s(x)}{\partial x^{\beta } } \right .
+ \frac{\sigma_m}{2 } m^2 \left [ s(x ) \right]^{\cal cpt } s(x ) \nonumber \\ & & \left . { } + \sigma_{\mu } { \mu}^2 \left\ { \left [ s(x ) \right]^{\cal cpt } s(x ) \right\}^2 \right ) -
\sigma_k \left [ \frac{\partial s(x)}{\partial x^{0 } } \right]^{\cal cpt } \frac{\partial s(x)}{\partial x^{0 } } \nonumber \\ & = & - \frac{\sigma_k } { 2 } \left [ \frac{\partial s(x)}{\partial x^{0 } } \right]^{\cal cpt } \frac{\partial s(x)}{\partial x^{0 } } - \frac{\sigma_k } { 2 } \left [ \nabla s(x ) \right]^{\cal cpt } \nabla s(x ) \nonumber \\ & & { } - \frac{\sigma_m}{2 } m^2 \left [ s(x ) \right]^{\cal cpt } s(x ) - \sigma_{\mu } { \mu}^2 \left\ { \left [ s(x ) \right]^{\cal cpt } s(x ) \right\}^2 \label{eq : ham}\end{aligned}\ ] ] all four terms in ( [ eq : ham ] ) include a ( positive definite ) @xmath26 norm .
by analogy with ordinary @xmath16 theory , the first two terms and the last must be positive if we are to have an energy spectrum that is bounded below .
we therefore set @xmath50 .
the sign of the @xmath44 term will be determined by the requirement of conformal invariance when we go to frw space .
the equation of motion for @xmath1 is obtained in the usual way , by requiring the action to be stationary under small variations , @xmath51 .
the appearance of @xmath52 in the action is unusual , and the variation requires some care ; details are given in the appendix .
the equation of motion is given in ( [ eqapp : motion ] ) .
writing @xmath50 , this becomes @xmath53 s(x ) = 0 \label{eq : motion}\ ] ] comparing this with mannheim s equation of motion , ( [ eq : s ] ) , we infer that we go over to frw space by setting @xmath54 .
since @xmath55 , we must set @xmath56 . to maintain mannheim s notation as far as possible
, we will here define @xmath57 , giving the equation of motion @xmath58 s(x ) = 0 \label{eq : motion2}\ ] ] following mannheim , we assume that @xmath1 develops a constant vacuum expectation value , calculated by setting the derivative term in ( [ eq : motion2 ] ) equal to zero .
the result is the same as ( [ eq : svev ] ) , @xmath59 , where , as before , @xmath55 and @xmath60
. we can now use ( [ eq : imink ] ) to write the action in frw space : @xmath61^{\cal cpt}_{;\mu } \left[s ( x ) \right]_{;\nu } \right .
\nonumber \\ & & \left .
{ } - \frac{r^{\sigma}_{\;\;\sigma}}{12 } s^{\cal cpt}(x ) s(x ) + \lambda_{\rm m } \left [ s^{\cal cpt}(x ) s(x ) \right]^2 \right\ } \label{eq : ifrw } \end{aligned}\ ] ] from this we can infer the energy - momentum tensor analogous to ( [ eq : t_mann ] ) . for a cosmological model comparable to mannheim
s only the last two terms are of interest , which are @xmath62 + g^{\mu \nu } \lambda_{m } \left [ s^{\cal cpt}(x ) s(x ) \right]^2 \label{eq : t_mann2}\ ] ] replacing @xmath32 by its vacuum expectation value , @xmath3 , we get @xmath63 the important point is that the signs of both these terms are reversed in comparison to ( [ eq : t_mann ] ) .
mannheim points out ( m6 , section 10 ) that because the frw space is conformally flat , the cosmological equation of motion ( cem ) reduces to @xmath64 @xmath65 is just the sum of ( [ eq : t_mann3 ] ) and @xmath66 , the contribution from ordinary matter ( fermion fields , electromagnetic radiation , etc . ) .
so our cem , analogous to m6 ( 222 ) , becomes @xmath67 or , as in m6 ( 223 ) @xmath68 with @xmath69 .
the @xmath4 term defines the effective gravitational constant in the theory .
we find @xmath70 analogous to m6 ( 224 ) , but wth a positive sign . of particular interest , since it permits an analytic solution , is a model containing radiation only ( m6 , section 10.2 ) .
a corresponding solution exists for the present model also ; let us see how it differs .
the equation analogous to m6 ( 226 ) is ( with @xmath71 ) : @xmath72 @xmath73 @xmath74 since @xmath75 for the open geometry of the mm , all three terms in ( [ eq : omega ] ) are positive , in contrast to m6 , where @xmath76 is negative .
write the solution as in m6 ( 230 ) , with @xmath77 : @xmath78 @xmath79 the first term on the right of ( [ eq : r ] ) is negative , not positive as in the mm .
this means that @xmath80 must start from zero , as in the standard cosmology , not from some finite value , as in the mm .
this is illustrated in figure [ fig : ccpgfigure1a_1b ] .
both curves were drawn using the formula ( [ eq : r ] ) , but in the upper graph @xmath81 , while in the lower graph @xmath82 .
the origin of @xmath83 is conventionally shifted in the lower graph so that @xmath84 is at point b , where @xmath85 .
two curves drawn using ( [ eq : r ] ) , but with different values of parameters .
( a ) @xmath86 , @xmath87 , @xmath88 .
@xmath80 starts from a non - zero value at @xmath84 ( point a ) .
( b ) @xmath86 , @xmath87 , @xmath89 .
part of the curve now lies below the horizontal axis , and is non - physical .
the origin of @xmath83 is conventionally shifted to b , where @xmath85 . ]
the present model has two features that are not present in conventional cosmology based on einstein s equations .
first , the model includes a scalar field that is essentially massless .
the non - zero vacuum expectation value of this field is essential , but we have ignored any excitations .
this may be permissible because the field couples very weakly to normal matter , and is difficult to observe , or it may undergo a spontaneous transition that renders it massive . second , mannheim uses ( [ eq : motion3 ] ) for his basic cosmological equation , rather than the more complete one that results from the conformal action , m6 ( 188 ) : @xmath90 where @xmath91 is the weyl tensor , defined in m6 ( 107 ) , ( 108 ) and ( 185 ) .
mannheim justifies the neglect of @xmath91 by observing that a frw metric is conformally flat , and in such a space @xmath92 .
this is true , but if we are to include perturbations to the metric ( as , for example , in the study of anisotropies of the cmb ) then this neglected term may become important .
we have pointed out two flaws in mannheim s conformal cosmological model .
@xmath2 the model predicts , apparently unambiguously , a negative value for the effective gravitational constant , @xmath0 .
the model involves a scalar field , @xmath36 , that satisfies a conformally invariant equation of motion and develops a vacuum expectation value , @xmath3 .
the values of the parameters that are needed to satisfy observations lead to a `` wrong sign @xmath22 theory '' , with a hamiltonian that has a spectrum that is unbounded below .
we have attempted to apply the techniques appropriate for such hamiltonians to this cosmological problem , restricting ourselves to the classical limit of field equations that are still imperfectly understood . in this limit , using assumptions that appear reasonable , we find both a positive value for @xmath0 and a positive definite spectrum for the hamiltonian .
our derivation depends on one simple observation , the change of sign as we go from ( [ eq : deltaika ] ) to ( [ eq : deltaikb ] ) .
the derivation presented here will remain conjectural until progress is made in two main directions : @xmath2 the techniques that have been successfully applied to @xmath16 quantum mechanics will have to be developed to cover the corresponding quantum field theory ; this seems not to have been achieved at this time @xcite .
in particular , the form ( [ eq : cptdef ] ) for @xmath33 must be shown to be appropriate at the classical level .
the various manipulations we have employed are suitable for minkowski space , but more detailed investigations are needed to show whether they can legitimately be extended to a frw space . if , on the other hand , we can accept the present model as viable , without first filling in these important gaps in our understanding , then we have to face the difficult question : how can we conclusively distinguish this model from the conventional one ? the department of physics at washington university , in particular carl bender , have given invaluable support to this retired colleague .
the action is defined in ( [ eq : imink ] ) .
we will start with the simplest term , @xmath93_{\forall \rho : z^{\rho } = -y^{\rho } } s(x ) \label{eq : ima}\end{aligned}\ ] ] let @xmath32 vary by a small amount @xmath94 .
then @xmath95 will vary by @xmath96 $ ] .
the variation of @xmath97 will be the sum of two terms , @xmath98 and @xmath99 : @xmath100_{\forall \rho : z^{\rho } = -y^{\rho } } \delta s(x ) \label{eq : deltaim1}\ ] ] @xmath101_{\forall \rho : z^{\rho } = -y^{\rho } } s(x ) \label{eq : deltaim2}\ ] ] take the complex conjugate of ( [ eq : deltaim2 ] ) , and let @xmath38 , @xmath39 : @xmath102 so that @xmath103 \nonumber \\ \fl & = & \int { \rmd}^4 x \int { \rmd}^4 y \ , ( -g)^{1/2 } \sigma_m m^2 c(x^{\mu } - y^{\mu } ) { \rm re } \left\{\left [ s^{*}(z ) \right]_{\forall \rho : z^{\rho } = -y^{\rho } } \delta s(x ) \right\ } \label{eq : deltaimb}\end{aligned}\ ] ] the @xmath104 term in the action can be treated in the same way , to give @xmath105 { \rm re } \left\ { \left [ s^ { * } ( z ) \right]_{\forall \rho : z^{\rho } = - y^{\rho } } \delta s ( x ) \right\}\end{aligned}\ ] ] calculating the variation of the kinetic term starts just as with conventional lagrangians .
we imagine a variation @xmath106 , and convert this to a variation of @xmath36 by an integration by parts , discarding a surface term . recalling that we are working in minkowski space , where the metric tensor is constant , we get @xmath107^{\cal cpt } \delta s ( x)\ ] ] using ( [ eq : cptdef ] ) we write this as @xmath108_{\forall \rho : z^{\rho } = -y^{\rho } } \delta s ( x ) \\
\fl & = & \int { \rmd}^4 x \int { \rmd}^4 y \ , ( -g)^{1/2 } \frac{\sigma_k } { 2 } g^{\mu \nu } \left [ \frac { \partial } { \partial y^{\nu } } c(x^{\mu } - y^{\mu } ) \right ] \nonumber \\ \fl & & \times \ , \left [ \frac{\partial s^ { * } ( z)}{\partial z^{\mu } } \right]_{\forall \rho : z^{\rho } = -y^{\rho } } \delta s ( x)\end{aligned}\ ] ] now integrate by parts and discard a surface term : @xmath109_{\forall \rho : z^{\rho } = -y^{\rho } } \delta s ( x ) \label{eq : deltaika } \\
\fl & = & \int { \rmd}^4 x \int { \rmd}^4 y \ , ( -g)^{1/2 } \frac{\sigma_k } { 2 } g^{\mu \nu } c(x^{\mu } - y^{\mu } ) \left [ \frac{\partial^2 s^ { * } ( z)}{\partial z^{\mu } \partial z^{\nu } } \right]_{\forall \rho : z^{\rho } = -y^{\rho } } \delta s ( x ) \label{eq : deltaikb } \end{aligned}\ ] ] the passage from ( [ eq : deltaika ] ) to ( [ eq : deltaikb ] ) is best illustrated by an example .
suppose @xmath110 where @xmath111 is some fixed vector .
then @xmath112_{\forall \rho : y^{\rho } = - x^{\rho } } = n ( -1)^{n } \left ( k_{\sigma } x^{\sigma } \right)^{n-1 } k_{\mu } k_{\nu } \label{eqapp : difex}\ ] ] and @xmath113_{\forall \rho : y^{\rho } = -x^{\rho } } & = & \left [ n \left ( k_{\sigma } y^{\sigma } \right)^{n-1 } k_{\mu } k_{\nu } \right]_{\forall \rho : y^{\rho } = - x^{\rho } } \nonumber \\ & = & n ( -1)^{n-1 } \left ( k_{\sigma } x^{\sigma } \right)^{n-1 } k_{\mu } k_{\nu } \label{eqapp : difex2}\end{aligned}\ ] ] note the change in sign from ( [ eqapp : difex ] ) to ( [ eqapp : difex2 ] ) ; it implies that the variation of the kinetic term results in @xmath114_{\forall \rho : z^{\rho } = -y^{\rho } } \delta s ( x ) \right\ } \label{eq : deltaikc}\ ] ] now use @xmath115 for arbitrary ( possibly complex ) variations @xmath116 to get the equation of motion @xmath117 s(x ) = 0 \label{eqapp : motion}\ ] ] | the conformal cosmological model presented by mannheim predicts a negative value for the effective gravitational constant , @xmath0 .
it also involves a scalar field , @xmath1 , which is treated classically . in this paper
we point out that a classical treatment of @xmath1 is inappropriate , because the hamiltonian is non - hermitean , and the theory must be developed in the way pioneered by bender and others .
when this is done , we arrive at a hamiltonian with an energy spectrum that is bounded below , and also a @xmath0 that is positive .
the resulting theory closely resembles the conventional cosmology based on einstein relativity . |
the standard model ( sm ) of particle physics has proven to hold up against any experimental tests it has been subjected to so far .
the sm has several striking features for which no underlying principle has been experimentally confirmed to this date .
first , the sm requires the scale of electro - weak symmetry breaking to be of order a few hundred gev , which is many orders of magnitude below the only fundamental scale of nature we know of , the planck scale .
second , to explain the masses and flavor violating transitions of fermions requires the fundamental yukawa matrices to satisfy a very particular scaling , for which no satisfactory symmetry or other underlying principle has been found so far .
while the scale of electro - weak symmetry breaking is known from the measured properties of gauge interactions , the scale of flavor violation could be completely unrelated to that scale .
however , many models of new physics which address the electro - weak scale also give additional contributions to flavor physics . thus , precise measurements of flavor and cp violating observables can severely constrain possible models of electro - weak symmetry breaking . since the standard model predicts the short distance couplings of quarks to one another , while experimental measurements are done with hadrons , one needs to understand long distance qcd effects on the measured quantities to extract the underlying physics .
it is the purpose of this talk to discuss how this separation between long and short distance physics can be achieved using effective theories .
the effective theory that is applicable to the non - leptonic @xmath0 decays to two light mesons , as we are concerned with here , is the soft - collinear effective theory ( scet ) @xcite .
there has been tremendous progress over the last few years in understanding charmless two - body , non - leptonic @xmath0 decays in the heavy quark limit of qcd @xcite . in this limit
one can prove factorization theorems of the matrix elements describing the strong dynamics of the decay into simpler structures such as light cone distribution amplitudes of the mesons and matrix elements describing a heavy to light transition @xcite .
it is very important that these results are obtained from a systematic expansion in powers of @xmath2 .
the development of soft - collinear effective theory ( scet ) @xcite allowed these decays to be treated in the framework of effective theories , clarifying the separation of scales in the problem , and allowing factorization to be generalized to all orders in @xmath3 .
factorization for @xmath4 decays involves three distinct distance scales @xmath5 . for @xmath4
decays , a factorization theorem was proposed by beneke , buchalla , neubert and sachrajda @xcite , often referred to as the qcdf result in the literature .
another proposal is a factorization formula which depends on transverse momenta , which is referred to as pqcd @xcite .
the factorization theorem derived using scet @xcite agrees with the structure of the qcdf proposal if perturbation theory is applied at the scales @xmath6 and @xmath7 .
one of the differences is that qcdf treats @xmath8 penguins perturbatively , while in the scet analysis they are left as a perturbative contribution plus an unfactorized large @xmath9 term .
the scet result improved the factorization formula by generalizing it to allow each of the scales @xmath6 , @xmath10 , and @xmath11 to be discussed independently .
the derivation of the scet factorization theorem occurs in several steps , corresponding to integrating out the various scales in the problem .
as already mentioned , the relevant scales are @xmath12 , @xmath13 and @xmath14 .
one starts from the effective weak hamiltonian , which describes the effects of the weak physics in terms of local 4-quark operators . integrating out @xmath15 fluctuations , the effective hamiltonian at leading order in @xcite
can be written schematically as @xmath16 \,,\end{aligned}\ ] ] where @xmath17 and @xmath18 are wilson coefficients , and the symbol @xmath19 denotes convolutions over various momentum fractions .
gluons are nonperturbative and lo soft gluons are exchanged by the @xmath20 , @xmath21 , @xmath22 and spectator quark which is not shown . , width=264 ] the term @xmath23 denotes operators appearing in long distance charm effects as in fig . [
fig : cpenguin ] .
there is broad agreement that charm loop contributions from hard ( @xmath24 ) momenta can be computed in perturbation theory and they are included in the @xmath17 and @xmath18 coefficients in eq .
( [ match ] ) .
however , there are non - perturbative contributions from penguin charm quark loops ( so - called charming penguins @xcite ) , which are contained in @xmath23 .
while no proof of factorization for the matrix element of this operator exists , it is still possible to determine its parametric dependence on @xmath25 , @xmath26 , and @xmath27 using operators in effective field theories .
we find @xmath28 thus , this contribution gives rise to a source of strong phases in the amplitudes , while all strong phases vanish for the other terms .
for this reason , we will keep this term and treat its matrix element as an unknown complex parameter in the theory .
the collinear fields in the operators @xmath29 can be decoupled from the ultrasoft fields by making a simple field redefinition @xcite .
the operators @xmath29 then factor into @xmath30 and @xmath31 parts , @xmath32 the first term contains both a soft @xmath20 quark field and a collinear field in the @xmath31 direction , while the second term contains two collinear fields in the @xmath33 direction .
the matrix element of the operators @xmath29 thus factor into a @xmath34 transition matrix element , and a vacuum to light meson matrix element . putting all these results together
, we obtain the scet factorization formula @xmath35 , \end{aligned}\ ] ] where @xmath36 and @xmath37 are non - perturbative parameters describing @xmath34 transition matrix elements , and @xmath38 parameterizes complex amplitudes from charm quark contractions for which factorization has not been proven .
power counting implies @xmath39 .
@xmath40 and @xmath41 are perturbatively calculable in an expansion in @xmath42 and depend upon the process of interest . at leading order in @xmath42 the short distance coefficients
@xmath40 is independent of the parameter @xmath43 , which implies that the functional form of the non - perturbative function @xmath44 does not matter , since we can define a new hadronic parameter @xmath45 .
without any theoretical input , there are 4 real hadronic parameters for each decay mode ( one complex amplitude for each ckm structure ) minus one overall strong phase . in addition , there are the weak cp violating phases that we want to determine . for @xmath46 decays
there are a total of 11 hadronic parameters , while in @xmath47 decays there are 15 hadronic parameters .
using isospin , the number of parameters is reduced .
isospin gives one amplitude relation for both the @xmath48 and the @xmath49 system , thus eliminating 4 hadronic parameters in each system ( two complex amplitudes for each ckm structure ) .
this leaves 7 hadronic parameters for @xmath46 and 11 for @xmath50 .
the su(3 ) flavor symmetry relates not only the decays @xmath46 and @xmath51 , @xmath52 , but also @xmath53 , @xmath54 and @xmath55 decays to two light mesons .
the decomposition of the amplitudes in terms of su(3 ) reduced matrix elements can be obtained from @xcite . and
20 hadronic parameters are required to describe all these decays minus 1 overall phase ( plus additional parameters for singlets and mixing to properly describe @xmath56 and @xmath57 ) .
of these hadronic parameters , only 15 are required to describe @xmath58 and @xmath47 decays ( 16 minus an overall phase ) .
if we add @xmath59 decays then 4 more paramaters are needed ( which are solely due to electroweak penguins ) . .
number of real hadronic parameters from different expansions in qcd .
the first column shows the number of theory inputs with no approximations , while the next columns show the number of parameters using only su(2 ) , using only su(3 ) , using su(2 ) and scet , and using su(3 ) with scet . for the cases with two numbers , @xmath60 , the second follows from the first after neglecting the small penguin coefficients , ie setting @xmath61 . in su(2 ) + scet @xmath62 has 6 parameters , but 1 appears already in @xmath63 , hence the @xmath64 .
the notation is analogous for the @xmath65 for @xmath66 .
[ table_parameters ] [ cols="<,^,^,^,^,^ " , ] [ example_table_2col ] and @xmath47 decays , with @xmath67 .
the 8 pieces of data in red ( below the dashed line ) have been used to determine the scet hadronic parameters @xmath68 , @xmath69 , @xmath70 as described in the text .
the data above the line are predictions .
the cp asymmetry in @xmath71 is expected to be small , but its numerical value is not predicted reliably .
[ fig2],width=321 ] experimentally , one finds @xmath72 and these results are consistent with zero .
the scet predictions are considerably more precise than the current measurements , and using conservative ranges @xmath73 , @xmath74 , @xmath75 , @xmath76 , @xmath77,and all phase differences @xmath78 one finds @xmath79 and for the cp - sum rule @xmath80 experimental deviations that are larger than these would be a signal for new physics . from table [ table_parameters ] and eq .
( [ params ] ) we can see that a total of 12 hadronic parameters are required to describe the decays @xmath46 , @xmath47 and @xmath59 at leading order in scet . on top of that , there is one weak phase @xmath81 which we will take as an unknown parameter .
however , in both the decays @xmath47 and @xmath82 , the coefficients multiplying the @xmath83 transition matrix elements @xmath84 and @xmath85 are very small , which implies that the observables are insensitive to the numerical value of these hadronic matrix elements .
this eliminates three of the hadronic parameters , namely @xmath86 , @xmath87 and @xmath88 .
finally , if we take the inverse moments of pion and kaon wave functions as experimental input , we obtain one additional relation between hadronic parameters .
this leaves us with a total of 7 hadronic parameters as well as one weak phase .
the 8 measurements used to fix these parameters are the branching ratios for @xmath0 decays to @xmath89 , @xmath90 , @xmath91 , @xmath92 , @xmath93 , as well as the cp asymmetries @xmath94 , @xmath95 and @xmath96 .
using the hadronic parameters extracted from the @xmath46 decays ( @xmath68 , @xmath69 and @xmath70 ) , the value for @xmath97 determined from the decays @xmath98 and @xmath99 decays and independently varying @xmath100 and @xmath101 , we can calculate all the remaining currently measured @xmath102 observables .
the results are shown in fig .
the data used in the fit are shown in red below the dashed dividing line while those above the line are predictions .
note that there is one more piece of data below the line than there are hadronic parameters .
this additional experimental information was used to determine the value @xmath103 .
we see that @xmath67 gives a good match to the @xmath104 data except for the asymmetry @xmath105 . when taking into account the theoretical error the most striking disagreements are the @xmath106 at @xmath107 and the cp - asymmetry @xmath108 at @xmath109 .
all other predictions agree within the uncertainties .
note that one could demand that @xmath108 be reproduced , which would imply a negative value of @xmath85 ( a naive fit for @xmath103 gives @xmath110 ) .
note however , that this would imply that both perturbation theory at the intermediate scale @xmath111 and su(3 ) are badly broken .
the above analysis has recently been repeated to include decays to iso - singlet final states @xcite .
this requires adding additional contributions which arise from purely gluonic configurations .
it turns out that the additional operators do not change the form of the factorization theorem given in eq .
( [ a0newfact ] ) , but the hadronic parameters @xmath112 , @xmath113 and @xmath114 receive order one contributions from these additional operators . to add isosinglet mesons to the phenomenological analysis thus requires a second set of parameters @xmath115 , @xmath116 and @xmath117 , which have to be determined from data separately .
since experimentally there are not enough decays available , su(3 ) flavor symmetry is required to retain predictive power . at the present time , there are two solutions possible for the gluonic hadronic parameters , and the degeneracy can only be lifted with further data .
the results of the global fit , as taken from @xcite , are shown in table [ tab1 ] .
in this talk i have discussed how one can separate the long distance non - perturbative physics from the underlying short distance physics using the soft - colinear effective theory .
one finds that the number of hadronic parameters is significantly reduced , such that they can be extracted directly from a subset of the data , and then used to make predictions for the remaining data .
i have given a brief discussion of the factorization theorem as it emerges from scet , and then discussed three phenomenological applications .
first , i gave a detailed counting of the hadronic parameters using various theoretical approaches , then i discussed a few impacts of the fact that there is only one source of strong phases in the decay amplitudes , and finally , i showed results for global analyses of @xmath0 decays to two pseudoscalar mesons , with and without including isosinglet mesons .
99 c. w. bauer , s. fleming and m. e. luke , phys .
d * 63 * , 014006 ( 2001 ) ; c. w. bauer , s. fleming , d. pirjol and i. w. stewart , phys .
d * 63 * , 114020 ( 2001 ) ; c. w. bauer and i. w. stewart , phys .
b * 516 * , 134 ( 2001 ) ; phys .
d * 65 * , 054022 ( 2002 ) ; c. w. bauer , s. fleming , d. pirjol , i. z. rothstein and i. w. stewart , phys .
d * 66 * , 014017 ( 2002 ) . m. beneke , g. buchalla , m. neubert and c. t. sachrajda , phys .
* 83 * , 1914 ( 1999 ) [ arxiv : hep - ph/9905312 ] ; nucl
. phys .
b * 591 * , 313 ( 2000 ) [ arxiv : hep - ph/0006124 ] ; m. beneke , g. buchalla , m. neubert and c. t. sachrajda , nucl . phys .
b * 606 * , 245 ( 2001 ) [ arxiv : hep - ph/0104110 ] ; m. beneke and m. neubert , nucl .
b * 675 * , 333 ( 2003 ) [ arxiv : hep - ph/0308039 ] .
keum et al , phys .
b * 504 * , 6 ( 2001 ) ; phys . rev .
d * 63 * , 054008 ( 2001 ) ; c. d. lu et al . , phys
d * 63 * , 074009 ( 2001 ) . c. w. bauer , d. pirjol , i. z. rothstein and i. w. stewart , phys
. rev .
d * 70 * , 054015 ( 2004 ) [ arxiv : hep - ph/0401188 ] . j. g. chay and c. kim , phys .
d * 68 * , 071502 ( 2003 ) [ arxiv : hep - ph/0301055 ] ; nucl .
b * 680 * , 302 ( 2004 ) [ arxiv : hep - ph/0301262 ] . c. w. bauer , d. pirjol and i. w. stewart , phys .
lett . * 87 * , 201806 ( 2001 ) [ arxiv : hep - ph/0107002 ] .
h. n. li , s. mishima and a. i. sanda , arxiv : hep - ph/0508041 . c. n. burrell and a. r. williamson , arxiv : hep - ph/0504024 .
m. ciuchini , e. franco , g. martinelli and l. silvestrini , nucl .
b * 501 * , 271 ( 1997 ) ; m. ciuchini et al .
b * 515 * , 33 ( 2001 ) .
m. ciuchini et al .
, arxiv : hep - ph/0208048 .
p. colangelo , g. nardulli , n. paver and riazuddin , z. phys .
c * 45 * , 575 ( 1990 ) .
d. zeppenfeld , z. phys .
c * 8 * , 77 ( 1981 ) .
m. j. savage and m. b. wise , phys .
d * 39 * , 3346 ( 1989 ) [ erratum - ibid .
d * 40 * , 3127 ( 1989 ) ] .
b. grinstein and r. f. lebed , phys .
d * 53 * , 6344 ( 1996 ) [ arxiv : hep - ph/9602218 ] .
a. r. williamson and j. zupan , arxiv : hep - ph/0601214 . | in this talk i will discuss non - leptonic @xmath0 decays , in particular how soft - collinear effective field theory ( scet ) can be used to constrain the non - perturbative hadronic parameters required to describe the various observables . |
is limited by noise in channels , but error correction methods can efficiently offset this restriction in both classical @xcite and quantum @xcite cases . at a simple level , multiple copies of the information can be transmitted , and a majority rule can be applied to discern the correct code , but such coding is neither practical nor efficient .
sparse graph coding , such as gallager s low - density parity - check ( ldpc ) codes , offers an efficient alternative that approaches the shannon information limit @xcite .
fortunately quantum coding and decoding strategies can be constructed from their classical counterparts , but unfortunately this mapping from classical to quantum coding can be problematic due to the requirement that quantum codes satisfy the duality - containing condition @xcite .
moreover , due to increased challenges posed by these quantum codes , performance improvement requires further progress in the proposed decoding algorithm @xcite .
the entanglement - assisted ( ea ) stabilizer formalism adds error - free entangled bits as a consumable resource for performing quantum error correction .
this ea approach overcomes the duality - containing requirement and thus offers a rich lode of quantum error correction protocols inspired by classical protocols @xcite . using the ea approach , modern codes for classical channels , such as sparse codes ,
can easily be imported as quantum error - correcting codes @xcite .
our aim is to improve belief propagation ( bp ) decoding methods so that quantum coding is dramatically improved over existing techniques .
specifically our numerical results presented here show that an improved bp method whose heuristical feedback strategies based on exploiting all accessible information from stabilizer measurements , yield a dramatically improved block error rate ( ber ) for any depolarizing channel .
our methods should work for any pauli noise channel . for @xmath0 qubits ,
a pauli channel is defined by the mapping @xmath1 with @xmath2 the @xmath0-fold tensor product of single - qubit pauli operators @xmath3 .
our interest is focused on memory - less channels wherein the error on each qubit is independent of the error on any other qubit .
in particular we consider the depolarizing channel , which is the most - studied case @xcite : for fixed channel error probability @xmath4 , the error on qubit @xmath5 is given by @xmath6 with @xmath7 denoting the probability of no error occurring . in quantum settings , due to the inability to measure each and every qubit
, syndrome - based decoding is typically chosen .
consequently , decoders using pauli channels are generally considered to be hard - decision decoders . in other words , in quantum decoding ,
the conventional soft - decision techniques are not applicable when the channel is in the pauli channel model .
generally , sparse quantum codes are decoded by using syndrome - based bp decoding algorithms which automatically imply hard - decision decoding .
though there is an equivalence between syndrome - based decoding and a posteriori probability decoding ( signal - based decoding ) under this setting , the syndrome - based decoding results in a serious drawback to bp decoder : the symmetric degeneracy error @xcite .
fortunately , poulin and chung ( pc08 ) @xcite propose a solution to the symmetric degeneracy error by using the random perturbation method .
motivated by the soft - decision techniques used in classical settings , which provide extra reliable information on the message nodes thereby yielding a better error correcting capacity @xcite , we develop a new heuristical feedback adjustment strategy for the standard bp decoder .
when used in decoding sparse quantum codes , our method can on one hand solve the symmetric degeneracy error problem , and on the other hand can provide more useful information to the message nodes .
the difference between pc08 and our approach is that , in pc08 , they feed back only the syndrome of the decoder output to adjust prior error probability distributions for received qubits .
these adjusted distributions are then fed back into the decoder .
we significantly improve this protocol by feeding back not just the syndrome but also the values of the frustrated checks on individual qubits of the code and the channel model ; accordingly we introduce a new adjustment strategy .
specifically , our approach , which is based not only on syndromes but also on frustrated checks obtained from full stabilizer measurements and the channel model , yields a better ber for the case of depolarizing quantum channels .
we provide a detailed description of our basic bp decoder for decoding sparse quantum codes .
the basic bp decoder introduced here is inspired by the strategy used for decoding sparse classical quaternary codes under the bp algorithm . using this strategy
, we can decode sparse quantum codes directly regardless of whether they arise from classical binary codes or not and regardless of whether they are calderbank - shor - steane ( css ) construction @xcite codes or not .
in this section , we briefly reprise the essential elements of bp iterative decoding for sparse quantum codes . in subsec .
[ subsec : bpdecoding ] we discuss the key idea of standard bp iterative algorithms .
then we compare decoding of classical codes vs quantum codes , and introduce the standard bp decoding for quantum codes . finally in subsec .
[ subsec : decodesparse ] we show how to decode sparse quantum codes in @xmath8 based on the standard bp algorithm .
consider a @xmath9-bit message that is encoded into an @xmath0-bit codeword , which is then transmitted through a noisy channel .
the received message is an error - prone vector over the output alphabet , and there is no guarantee that decoding will reveal the original codeword .
however , the codeword can be guessed with a high probability of being correct by maximizing the probability for this codeword based on the observed output vector @xcite .
unfortunately for a linear block code , encoding @xmath9 information bits into @xmath0 bits allows @xmath10 possible @xmath0-bit codewords , and calculating conditional probabilities for individual codewords is too expensive in practice .
bp algorithms overcome this inefficiency for sparse codes @xcite .
the strategy is to represent a linear - block classical error - correcting code ( cecc ) by a tanner graph comprising message nodes and check nodes corresponding respectively to received bits and check constraints .
then an iterative algorithm recovers the values of received bits as follows @xcite . at each round
, probabilities are passed from message nodes @xmath11 to check nodes @xmath12 and then from @xmath12 back to @xmath11 .
the probabilities from @xmath11 to @xmath12 are computed based on the observed value of the message node and other probabilities passed from the neighboring check nodes to the respective message node .
it turns out that , for decoding sparse codes , the bp algorithm can provide a reasonable trade - off between complexity and performance . for a linear block cecc
, the code space can be viewed as the orthogonal projection space ( solution space ) of its check matrix . the sender alice transmits her message as a codeword that is encoded according to a specific check ( or generator ) matrix through the channel .
when receiver bob obtains the channel output , the received vector may not be the solution vector of the check matrix .
therefore bob needs to apply a smart algorithm to recover the codeword efficiently . as
the check matrix is sparse , bob employs the bp algorithm .
first , bob measures each bit to obtain its posterior probability distribution .
subsequently , he puts these probabilities into the bp decoder , and , based on the constraint that the codeword should be the orthogonal vector of the check matrix , he infers the original message .
this procedure extends naturally to sparse quantum codes . in the quantum case ,
the stabilizer formalism for a quantum error - correcting code ( qecc ) is useful .
the code space within @xmath13 corresponds to the simultaneous @xmath14 eigenspace of all @xmath15 generators of an abelian subgroup @xcite @xmath16 alice transmits her message as a codeword , which propagates through the depolarizing quantum channel .
due to channel errors , the received codeword may not be the simultaneous @xmath14 eigenspace of @xmath17
. bob measures @xmath17 to obtain the syndrome=0 $ ] , @xmath18 or else @xmath19 , where @xmath20 and @xmath21 can be viewed as a check node in the tanner graph of a qecc with stabilizer @xmath17 from eq . (
[ eq : stabilizer ] ) . ] @xmath22 and then estimates which error @xmath23 from eq .
( [ eq : paulinoisechannel ] ) has occurred based on the noise model and the syndrome , which is effectively a check constraint .
bob redresses the error by applying the same error operation he inferred onto his received codeword to try to recover the quantum code . for pauli channels all errors are @xmath0-fold tensor products of single - qubit pauli operators
hence square to the identity ; thus re - applying the same error should restore the original codeword . due to the close analogy between ceccs and qeccs as discussed above , it is tempting to extend the bp iterative decoding algorithm to quantum settings .
in fact a qecc can also be represented by a decorated tanner graph @xcite .
therefore , it is natural to transmit the messages from @xmath12 to @xmath11 and the messages from the @xmath11 to @xmath12 according to the principles of bp algorithms @xcite .
here we first briefly review the bp algorithm used for decoding sparse quantum codes @xcite but begin with necessary terminology .
the probability of pauli error @xmath24 occurring on qubit @xmath5 is @xmath25 .
the neighborhood of qubit @xmath5 is denoted by @xmath26 , and we define @xmath27 similarly as the set of qubits connected to check @xmath21 . the tensor product of pauli errors over @xmath27 is denoted @xmath28 . then , following pc08 s notation , for two @xmath29-qubit case , @xmath30 means @xmath31=0 $ ] and @xmath32 means @xmath33 . for two @xmath0-qubit pauli operators such as @xmath34 and @xmath35
, we have @xmath36 the messages from check nodes to qubit nodes are denoted @xmath37 which is defined only up to a constant factor .
the factor can be fixed by normalization @xcite .
the messages from qubit nodes to check nodes are similarly denoted as @xmath38 where @xmath25 is the initial probability in the definition of the memoryless channel .
then the beliefs @xmath39 are constructed by first initializing @xmath40 , then evaluating according to @xmath41 after the iteration procedure based on eqs .
( [ eq : bpequation1 ] ) and ( [ eq : bpequation2 ] ) .
in fact , eqs .
( [ eq : bpequation1 ] ) and ( [ eq : bpequation2 ] ) define a sum - product iterative procedure for decoding sparse quantum codes .
hence this algorithm is also called the sum - product algorithm ( spa ) , which is one of the most important algorithms based on bp . in order to clarify the bp decoding algorithm for sparse quantum codes
, we now show how to implement eqs .
( [ eq : bpequation1 ] ) and ( [ eq : bpequation2 ] ) in gf(4 ) .
there is a convenient isomorphism between the pauli group @xmath42 generated by @xmath43 and the galois field @xmath8 generated by @xmath44 .
the isomorphism is explained by the element identification @xmath45 and the operation identification _ _ multiplication__@xmath46_addition _ and _ _
commutativity__@xmath46_trace inner product _ @xcite . for the one - qubit case
, we have , @xmath47&=0 \leftrightarrow { \rm tr}(\hat{p } \times \bar{\hat{q } } ) = 0 , \ ; \ { p , q \}=0 \leftrightarrow { \rm tr}(\hat{p } \times \bar{\hat{q } } ) = 1\end{aligned}\ ] ] the isomorphism is readily extended to the @xmath0-qubit case : @xmath48&=0 \leftrightarrow { \rm tr}(\bm{u}_p \cdot \bm{v}_q ) = 0,\ ; \ { p , q \}=0 \leftrightarrow { \rm tr}(\bm{u}_p
\cdot \bm{v}_q ) = 1.\end{aligned}\ ] ] where ` @xmath49 ' used here ( between two vectors ) is a regular inner product .
that is , for @xmath50 and @xmath51 , we have , @xmath52 the addition and multiplication rules of @xmath8 are shown in tables [ table : isomorphictable1 ] and [ table : isomorphictable2 ] , respectively .
.addition of gf(4 ) [ cols="^,^,^,^,^",options="header " , ] we can import the strategy used for decoding sparse classical quaternary codes under bp to use for decoding sparse quantum codes .
this adaptation to quantum codes is achieved by transforming the check criterion ( [ eq : bpequation1 ] ) from commutativity to trace inner product .
for example , suppose @xmath53 .
then the check criterion for this check node should be @xmath54 , which is equivalent to @xmath55=0 $ ] . according to isomorphism ( [ eq : isomorphic2 ] ) , @xmath56 , which implies that either @xmath57 or @xmath58 because @xmath59 and @xmath60 .
we define @xmath61 $ ] as the probability for @xmath62 to take the value @xmath63 , with @xmath64 the @xmath65 entry of @xmath66 , and @xmath67 the mapped elements of @xmath24 and @xmath68 in gf(4 ) , see eq .
( [ eq : isomorphic1 ] ) . ] and @xmath69 $ ] the probability of @xmath70 ( excluding the @xmath65 entry ) to take the value @xmath63 .
as @xmath71 , @xmath72 thus , @xmath73&=\left(p[0]+p[1]\right)/2=p_q[1 ] , \\ \label{eq : communicatecase2 } p_q[\omega]&=\left(p[\omega]+p[\bar{\omega}]\right)/2=p_q[\bar{\omega}].\end{aligned}\ ] ] similarly , if @xmath19 , then @xmath74 , which implies that either @xmath75 or @xmath76 because @xmath77 and @xmath78 .
according to table [ table : isomorphictable1 ] , @xmath79=\left(p[\omega]+p[\bar{\omega}]\right)/2=p_q[1],\ ] ] @xmath80=\left(p[0]+p[1]\right)/2=p_q[\bar{\omega}].\ ] ] with this knowledge , the message @xmath81 can be computed by : @xmath82.\ ] ] given @xmath83 , @xmath84 can then be derived directly by substituting eq .
( [ eq : computemessage ] ) into eq .
( [ eq : bpequation2 ] ) .
decoding based on @xmath8 affords the advantage that sparse quantum codes can be decoded without losing correlations between errors that would otherwise impact the coding scheme s performance . in particular sparse quantum codes
can be decoded directly regardless of whether they arise from a classical binary codes or not and regardless of whether they follow a css construction code or not . for comparison
, we henceforth refer to the decoder based on this bp decoding algorithm as the ` standard bp decoder ' .
in 2008 pc08 proposed a random perturbation strategy based on the syndrome of the bp decoder output to overcome the symmetric degeneracy problem .
the random perturbation method is simple and efficient but unfortunately overlooks many errors .
as pointed out in pc08 , all errors in their simulations could be attributed to the decoder rather than to the finite minimal distance of the code .
we observe the same problems as pc08 do when using the random perturbation method .
in particular , most errors in the simulations are _ detected _ errors , not the undetected errors .
this result suggests that feeding back useful information to the standard bp decode could help .
the feedback would help the bp decoder determine a valid output @xmath85 , whose syndrome @xmath86 is identical with the observed syndrome @xmath87
. then @xmath85 will equal @xmath23 with high probability , for @xmath23 the error occurring on the transmitted quantum state during transmission . in this section
, we present our enhanced feedback bp iterative decoding algorithm , which provides useful information to the bp decoder based on exploiting not only the syndrome but also the stabilizer itself and the channel model in the feedback procedure .
then suppose that , subsequent to the standard bp decoding procedure , @xmath86 is not identical with the observed syndrome @xmath87
. then evidently at least one entry of @xmath85 occurring on one of the qubits connected to check @xmath66 has an error , for @xmath66 a frustrated check .
if we can reset the initial probability distributions for errors occurring on the qubits connected to @xmath66 in a more reliable way , but not simply by adopting the prior probability distributions obtained from the channel model , more useful information could be provided to the bp decoder .
furthermore symmetric degeneracy could also be overcome as well .
thus it is reasonable to expect that bp decoding ability will be significantly improved .
our new feedback bp iterative decoder is devised as follows .
first the standard bp decoder is used to identify the error @xmath23 based on the observed syndrome and the prior probability distributions of each entry of @xmath23 .
when the standard bp decoder fails ( @xmath88 ) , we randomly choose a frustrated check , say @xmath66 .
there are only two results for this frustrated check : @xmath89 and @xmath90 or else @xmath53 and @xmath91 .
either way we choose a random qubit that connects to @xmath66 , for example qubit @xmath5 , and then reset @xmath25 according to the value of @xmath64 ( the entry to which qubit @xmath5 corresponds in @xmath92 ) and the channel model as follows .
let @xmath93 be three distinct elements .
suppose @xmath94 , and the channel is the depolarizing channel with crossover probability @xmath95 , for @xmath96 established in eq .
( [ eq : depolarizingchannel ] ) . then ,
if @xmath97 and @xmath98 , we reset @xmath99 for example , if @xmath100 , we obtain @xmath101 instead suppose @xmath102 and @xmath103 . then we reset @xmath104 in this case , if @xmath100 , we obtain @xmath105 next , following a similar strategy to pc08 s , we feed back this adjustment to the standard bp decoder and let it iterate with this modified probability distribution for @xmath106 ( a pre - determined fixed number of the reiteration ) steps . if bp halts during this period ( i.e. @xmath86 corresponds to the syndrome @xmath87 ) , the procedure is complete and terminates .
if @xmath107 is still frustrated , we restore the probability distribution @xmath25 and choose a different qubit connected to @xmath107 , say qubit @xmath108 , and reset @xmath109 based on the same strategy that has already been used for resetting @xmath25 .
if @xmath21 is not frustrated , but the halting condition is not yet satisfied , we choose another frustrated check @xmath110 and adjust the probability distribution of the error occurring on one of its qubits as we do for a frustrated check @xmath21 .
now suppose that the decoder still can not yield a valid output after trying a pre - determined number @xmath111 feedback adjustments for qubits according to the procedures mentioned above .
in other words , suppose that , after traversing @xmath111 entries associated with frustrated checks , no valid result is yielded by the decoder . in this case , we allow the decoder to terminate this procedure and report a failure . without the termination condition , the decoder is wasting time traversing all entries associated with frustrated checks and thereby introducing unwanted redundancy into the procedure .
( [ eq : resetting1 ] ) and ( [ eq : resetting2 ] ) are justified for the following reason .
empirically , the decoder is naturally too biased towards the identity @xmath112 @xcite due to the prior probability distribution obtained from the channel model .
this biased probability distribution will lead to @xmath113 while @xmath114 .
heuristically , we reset @xmath25 to make it anti - commute with @xmath115 with a greater probability , since at least one entry of @xmath23 anti - commutes with its corresponding entry in @xmath92 .
the channel is depolarizing that is pauli errors @xmath116 , @xmath117 , @xmath118 occur with equal probability , so , in eq .
( [ eq : resetting1 ] ) , we let @xmath119 and @xmath120 share the probability as @xmath112 . , we let @xmath112 and @xmath121 share @xmath122 equally for simplicity . ]
when @xmath102 , setting @xmath123 to @xmath112 is conducive to obtaining @xmath124 .
another pauli operator could also be conducive to making @xmath124 but is ignored by the prior probability distribution .
we can readily ascertain this pauli operator according to the entries of @xmath125 . according to eq .
( [ eq : resetting2 ] ) , we let this pauli operator ( @xmath121 ) share the same probability @xmath96 with @xmath112 and have @xmath119 and @xmath120 share the small probability @xmath126 equally .
our strategy retains pc08 s capability of breaking the symmetric degeneracy while feeding back more useful information to the bp decoder than pc08 in order to help the decoder determine a valid output .
more useful information comes from retaining full information about the entire stabilizer rather than just a syndrome .
specifically , our feedback adjustment strategy is not solely based on the syndrome but also on the channel model and on the individual values of the entries of the frustrated checks ( the checks are just the generators of the stabilizer ) .
in this section we illustrate our enhanced feedback iterative decoding algorithm to show how it succeeds in helping the bp decoder find a valid output . for @xmath127 $ ] denoting an ea qecc that encodes @xmath9 qubits into @xmath0 qubits with the help of @xmath107 ancillary ebits @xcite
, we construct a simple ea qecc for @xmath128 $ ] as follows . according to the ea qecc formalism , we start with a classical @xmath129 $ ] quaternary code with check matrix @xmath130 first we transform @xmath131 to @xmath132 then transform @xmath133 to a set of ( perhaps non - commuting ) generators @xmath134 next we transform @xmath135 into a canonical form @xmath136 by multiplying the third generator by the second generator and by multiplying the fourth generator by the first and second generator .
thus , we obtain @xmath137 now , it is easy to check that the first generator anti - commutes with the second generator , and the last two generators commute with each other as well as commuting with the first two generators . according to the ea qecc formalism
, this coding scheme needs one ebit to assist encoding .
the extended commuting set of generators is @xmath138 for @xmath139 which is just the stabilizer of @xmath140 $ ] .
it is easy to check that @xmath141 is isomorphic to @xmath142 which means we can encode one logical qubit into four physical bits with the help of one entangled pair plus two ancilla bits via a unitary transformation @xcite . as the half of the entangled pair being held by the receiver
is assumed to be error - free , the error - correcting capacity of @xmath140 $ ] only depends on the left part of @xmath141 , which corresponds to the four transmitted qubits held by the sender .
suppose the qubits of @xmath128 $ ] are sent through a depolarizing channel and error @xmath143 is applied to the four transmitted qubits .
measurement of the stabilizer reveals the error syndrome @xmath144 . in fig .
[ fig : standardbp ] , we display the performance of the standard bp decoder under this setting . for an ea qecc with stabilizer generators
@xmath145 , @xmath146 , @xmath147 and @xmath148 and with syndrome @xmath149 using the standard bp decoding algorithm .
the prior @xmath25 of the four sent qubits are obtained from the depolarizing channel with @xmath150 .
( @xmath151 ) , ( @xmath152 ) , ( @xmath153 ) and ( @xmath154 ) show the beliefs for qubits @xmath29 , @xmath155 , @xmath156 and @xmath157 , respectively . qubit @xmath158 is assumed to be held by the receiver and thus to be error - free .
the symbols @xmath159 , @xmath160 , @xmath161 and @xmath162 indicate the beliefs for @xmath112 , @xmath116 , @xmath117 and @xmath118 , respectively . in this case , the output of the standard bp decoder is @xmath163 , which is a detected error . , width=170 ] for an ea qecc with stabilizer generators @xmath145 , @xmath146 , @xmath147 and @xmath148 and with syndrome @xmath149 using the standard bp decoding algorithm .
the prior @xmath25 of the four sent qubits are obtained from the depolarizing channel with @xmath150 .
( @xmath151 ) , ( @xmath152 ) , ( @xmath153 ) and ( @xmath154 ) show the beliefs for qubits @xmath29 , @xmath155 , @xmath156 and @xmath157 , respectively . qubit @xmath158 is assumed to be held by the receiver and thus to be error - free .
the symbols @xmath159 , @xmath160 , @xmath161 and @xmath162 indicate the beliefs for @xmath112 , @xmath116 , @xmath117 and @xmath118 , respectively . in this case , the output of the standard bp decoder is @xmath163 , which is a detected error .
, width=172 ] for an ea qecc with stabilizer generators @xmath145 , @xmath146 , @xmath147 and @xmath148 and with syndrome @xmath149 using the standard bp decoding algorithm
. the prior @xmath25 of the four sent qubits are obtained from the depolarizing channel with @xmath150 .
( @xmath151 ) , ( @xmath152 ) , ( @xmath153 ) and ( @xmath154 ) show the beliefs for qubits @xmath29 , @xmath155 , @xmath156 and @xmath157 , respectively . qubit @xmath158 is assumed to be held by the receiver and thus to be error - free .
the symbols @xmath159 , @xmath160 , @xmath161 and @xmath162 indicate the beliefs for @xmath112 , @xmath116 , @xmath117 and @xmath118 , respectively . in this case , the output of the standard bp decoder is @xmath163 , which is a detected error . , width=170 ] for an ea qecc with stabilizer generators @xmath145 , @xmath146 , @xmath147 and @xmath148 and with syndrome @xmath149 using the standard bp decoding algorithm .
the prior @xmath25 of the four sent qubits are obtained from the depolarizing channel with @xmath150 .
( @xmath151 ) , ( @xmath152 ) , ( @xmath153 ) and ( @xmath154 ) show the beliefs for qubits @xmath29 , @xmath155 , @xmath156 and @xmath157 , respectively . qubit @xmath158 is assumed to be held by the receiver and thus to be error - free .
the symbols @xmath159 , @xmath160 , @xmath161 and @xmath162 indicate the beliefs for @xmath112 , @xmath116 , @xmath117 and @xmath118 , respectively . in this case , the output of the standard bp decoder is @xmath163 , which is a detected error .
, width=172 ] from fig .
[ fig : standardbp ] , it is easy to check that the syndrome of the output of the standard bp decoder is @xmath164.]$ ] is designed to correct one error but not two errors .
] we now use the standard bp decoder with pc08 s random perturbation . comparing between the error syndrome @xmath144 and the syndrome of the output of the standard bp decoder @xmath164
, the probability distributions of the errors occurring on the qubits connected to the second , third or fourth checks can be reset .
for all these cases , the decoder could not yield an appropriate recovery . here as an example , we use pc08 s random perturbation strategy for the frustrated check @xmath165 and show the corresponding performance in fig . [
fig : syndrome - basedbp ] . but using the standard bp decoding algorithm replaced by pc08 and applying random perturbation strength @xmath166 to the prior for qubits @xmath29 , @xmath155 and @xmath157 .
, width=169 ] but using the standard bp decoding algorithm replaced by pc08 and applying random perturbation strength @xmath166 to the prior for qubits @xmath29 , @xmath155 and @xmath157 . ,
width=172 ] but using the standard bp decoding algorithm replaced by pc08 and applying random perturbation strength @xmath166 to the prior for qubits @xmath29 , @xmath155 and @xmath157 .
, width=169 ] but using the standard bp decoding algorithm replaced by pc08 and applying random perturbation strength @xmath166 to the prior for qubits @xmath29 , @xmath155 and @xmath157 . ,
width=172 ] finally , we use our enhanced feedback bp decoding algorithm . by the same method ,
our feedback strategy can reset the probability distributions of the errors occurring on the qubits connected to the second , third or the fourth checks .
when the decoder chooses the fourth entry ( @xmath116 ) of the frustrated check @xmath165 and resets @xmath167 according to eq .
( [ eq : resetting2 ] ) , the correct decoding result arises in just a few iterations .
we show the performance of our approach in fig .
[ fig : stabilizer - basedbp ] . but using our enhanced feedback iterative decoding algorithm . here
we reset @xmath167 according to eq .
( [ eq : resetting2 ] ) . in this case
, only three iterations were required to yield a valid output @xmath168 , which is exactly the error occurring on [ [ 4 , 1;1 ] ] during its transmission . ,
width=170 ] but using our enhanced feedback iterative decoding algorithm . here
we reset @xmath167 according to eq .
( [ eq : resetting2 ] ) . in this case
, only three iterations were required to yield a valid output @xmath168 , which is exactly the error occurring on [ [ 4 , 1;1 ] ] during its transmission .
, width=172 ] but using our enhanced feedback iterative decoding algorithm . here
we reset @xmath167 according to eq .
( [ eq : resetting2 ] ) . in this case
, only three iterations were required to yield a valid output @xmath168 , which is exactly the error occurring on [ [ 4 , 1;1 ] ] during its transmission . ,
width=170 ] but using our enhanced feedback iterative decoding algorithm . here
we reset @xmath167 according to eq .
( [ eq : resetting2 ] ) . in this case
, only three iterations were required to yield a valid output @xmath168 , which is exactly the error occurring on [ [ 4 , 1;1 ] ] during its transmission .
, width=172 ] in this example , pc08 s method could not help the standard bp decoder to yield a valid output , however , our enhanced feedback bp decoding algorithm yields a valid output in just a few iterations .
in fact even if the error occurring on the first four qubits is not @xmath169 but @xmath170 , which has the same syndrome @xmath144 as @xmath169 , our decoding output @xmath169 can still recover the transmitted quantum state because @xmath171 which is just the third generator of the extended stabilizer . because it is hard to check whether @xmath172 when the number of generators of @xmath17 is large , we choose @xmath173 as the success criterion of the decoding result in our simulations .
we have applied our enhanced feedback iterative bp decoding algorithm to a variety of sparse quantum codes , including conventional sparse quantum codes and ea sparse quantum codes over depolarizing channels . in each case , our improved bp decoder yields significantly lower ber over both the standard bp decoder and the bp decoder with pc08 s random perturbation . in the following subsections , we simulate decoding of sparse quantum codes with different code parameters ( including block lengths , rates , and row weight ) under the three decoders to demonstrate the superiority of our approach . conventional sparse quantum codes can be constructed from sparse classical dual - containing codes .
one of the most successful dual - containing constructions is the so - called `` construction b '' @xcite , which is built as follows .
first we take an @xmath174 cyclic matrix @xmath153 with row weight @xmath175 , and define @xmath176.\ ] ] then we delete some rows from @xmath177 to obtain a matrix @xmath178 with @xmath15 rows . by construction , @xmath178 is dual - containing . a conventional sparse quantum code with length @xmath0 can thus be given according to css construction .
our first example is based on this strategy .
we first construct a cyclic binary ldpc code [ 63 , 37 ] with row weight @xmath179 based on finite geometries @xcite .
this code has a @xmath180 cyclic check matrix with @xmath181 independent rows and @xmath182 redundant rows .
hence , we can construct a conventional sparse quantum code with @xmath183 , @xmath184 , and @xmath185 .
we refer to this @xmath186 $ ] quantum ldpc code with rate @xmath187 and row weight @xmath188 as conventional " .
we show the performance of the standard bp decoder , the bp decoder with pc08 s random perturbation and our enhanced feedback bp decoder when applied to conventional over depolarizing channels in fig .
[ fig : conventionalexample ] . for three different bp decoders when applied to conventional .
the maximum number of iterations is @xmath189 .
the maximal number of iterations between each perturbation is @xmath190 .
the strength of the random perturbation used here is @xmath191 .
the maximal number of entries traversed by enhanced feedback decoder round down to one fifth of the code length , which equals 25 here.,width=283 ] ea sparse quantum codes can be constructed from arbitrary binary or quaternary sparse classical codes @xcite . here
we present the decoding of three ea sparse quantum codes that are constructed in the same way as the construction of @xmath192 $ ] discussed in sec .
[ sec : simpleexample ] .
these three codes have different net rate ( including positive , zero and negative net rates ) , different block lengths and different row weight distributions .
the first ea sparse quantum code we consider is @xmath193 $ ] , which is constructed from an irregular quaternary sparse classical code @xmath194 $ ] with girth @xmath195 .
the classical code @xmath196 $ ] is designed by using the progressive - edge - growth ( peg ) algorithm under the constraint that the degree distribution of bit nodes is @xcite @xmath197 as there are @xmath198 pairs of anti - commuting generators in @xmath136 according to the construction method of eq .
( [ eq : canonicalstabilizer ] ) , then the ea formalism yields the ea sparse quantum code @xmath199=[[128 , 2\times 79 - 128 + 49 ; 49]],\ ] ] which can be constructed from the classical quaternary code @xmath200=[128 , 79]$ ] @xcite
. the code @xmath201 $ ] has a positive net rate @xmath202 here we refer to this code as ea-1 " and show the performance of the three decoders with respect to ea-1 over depolarizing channels .
the performance , in terms of block error rate , is summarized in fig .
[ fig : eaexample1 ] .
our second ea sparse quantum code is @xmath203 $ ] , which is constructed from a binary regular sparse classical code @xmath204 $ ] .
the classical @xmath205 $ ] , which has a rate @xmath206 , row weight @xmath207 and column weight @xmath156 , was constructed by mackay @xcite .
as there are @xmath208 pairs of anti - commuting generators and @xmath179 commuting generators in @xmath136 , we obtain an ea sparse quantum code @xmath203 $ ] according to the ea stabilizer formalism . here
we take this code as our second example of ea sparse quantum codes whose net rate is @xmath209 and call it ea-2 " .
we show the performance of the three decoders when applied to ea-2 over depolarizing channels in fig .
[ fig : eaexample2 ] in terms of ber .
the third ea sparse quantum code is @xmath210 $ ] , which is constructed from @xmath211 $ ] , a famous sparse classical code proposed as the international standard for cellular telephones .
the @xmath211 $ ] code , which has a rate @xmath212 and row weight @xmath157 , was also constructed by mackay @xcite . following the construction procedure mentioned in sec .
[ sec : simpleexample ] , the number of anti - commuting pairs in @xmath136 is @xmath213 ; thus , we obtain an ea sparse quantum code @xmath210 $ ] . here
we take this code as our third example of ea sparse quantum codes , this time with a negative net rate @xmath214 we name it ea-3 " and show the performance of the three bp decoders when applied to ea-3 over depolarizing channels in fig .
[ fig : eaexample3 ] .
we evaluate and compare the time consumed by each decoder according to the average number of iterations @xmath215 consumed by one codeword , with @xmath216 the total number of iterations to finish decoding @xmath217 blocks . in our simulation @xmath218 . in figs . 8 and 9 we show the time efficiency ( complexity ) comparison for the three decoders applied to conventional , ea-1 , ea-2 and ea-3 .
evidently our algorithm has a smaller anoi than pc08 , which shows improved decoding efficiency when compared to pc08 . from figs .
[ fig : conventionalexample]-[fig : averageiteration2 ] , it is evident that our strategy yields a significantly lower ber with a lower anoi for both conventional sparse quantum codes and ea sparse quantum codes . as an example of the effectiveness of our new decoder ,
consider the case @xmath219 , fig .
4 shows that the enhanced - feedback bp decoder yields a @xmath207 db gain over pc08 and an @xmath179 db gain over the standard bp decoder while reducing the anoi from 3.1 for pc08 to 2.2 for our enhanced - feedback bp decoder as seen in fig .
8 . for another point of view , in fig .
[ fig : eaexample1 ] , we can see that to guarantee ea-1 having a decoding error below @xmath220 , the standard bp decoder allows @xmath221 and pc08 allows @xmath222 .
this value is increased dramatically to @xmath223 using our enhanced feedback bp decoder with a similar anoi performance as pc08 . as for the decoding of ea-2 ( fig .
[ fig : eaexample2 ] ) , when @xmath224 , the ber performance of pc08 s method is @xmath225 times better than that of the standard bp decoder with @xmath226 times more anoi , and the ber performance of our approach is approximately @xmath227 times better than that of the standard bp decoder with only @xmath228 times more anoi .
notably , interesting results appear in the decoding of ea-3 . from fig . [
fig : eaexample3 ] and [ fig : averageiteration2 ] we see that pc08 could not significantly improve the standard bp decoder even with high anoi , yet our approach does improve the performance of ea-3 significantly with almost the same anoi as the standard bp decoder excepting the last two nodes .
our algorithm is capable of turning the detectable errors into new outputs whose syndromes are identical with corresponding observed syndromes . we refer to these outputs as _ valid _ outputs can be classified into three cases : ( 1 ) @xmath229 ; ( 2 ) @xmath230 but @xmath231 ; ( 3 ) @xmath232 .
the first two cases will yield correct outputs that can recover the sent quantum states successfully .
however , it is hard to check whether @xmath231 which makes distinguishing cases ( 2 ) and ( 3 ) difficult . therefore , we choose ( 1 ) as the success criterion for the decoding result in our numerical simulation .
some detectable errors are turned into undetectable errors by our decoder . in order to show how often our algorithm results in an undetectable error
, we classify all the results yielded by our algorithm ( not just valid outputs ) when the original bp protocol fails into three cases : ( 1 ) ; cases ( 2 ) or ( 3 ) ; remain as detectable errors . in fig .
[ fig : conventional - proportion ] , we take conventional as an example to depict the number of blocks that fall into each whereas the original bp protocol fails 50 times . for our enhanced bp decoder when applied to conventional while the original bp decoder fails 50 times .
the maximum number of iterations is @xmath189 .
the maximal number of iterations between each perturbation is @xmath190 .
the maximal number of entries traversed by enhanced feedback decoder round down to one fifth of the code length.,width=283 ] .
[ fig : conventional - proportion ] it is evident to see from fig .
[ fig : conventional - proportion ] that most of the detectable errors will be turned to correct output but not undetectable errors by using our enhanced feedback bp decoder .
it is evident from fig .
[ fig : conventional - proportion ] that most of the detected errors will be turned to correct output but not un - detected errors by using our enhanced feedback bp decoder .
in fact we can take the percentage of detected ( pod ) errors as an important parameter to appraise the decoding capacity of the decoder , since one detected error indicates a failure of the decoder and undetected errors can be viewed as fundamental limitations of the code in some sense . here
we take the ea-1 as an example and suppose this time @xmath233 . in this case
our simulation reveals that standard bp decoder fails @xmath234 times when it has simulated @xmath235 blocks , and the number of detected error is @xmath198 so pod@xmath236 .
similarly pc08 fails @xmath234 times when it has simulated @xmath237 blocks , and the number of detected errors is also @xmath238 so again pod@xmath236 .
on the other hand , our enhanced feedback bp decoder fails @xmath234 times for a simulation of @xmath239 blocks , and only @xmath240 errors are attributed to detected errors so pod=@xmath241 .
the pod in our case is much less than for the other two bp decoders .
we have developed an enhanced feedback bp decoding algorithm whose feedback adjustment strategy is not based solely on the syndrome but also on the channel model and on the individual values of the entries of the frustrated checks .
our approach retains the capability of breaking the symmetric degeneracy while also feeding back extra useful information to the bp decoder .
therefore , our feedback adjustment strategy yields a better error correcting capacity with relative less iterations compared to the proposed decoding algorithms for sparse quantum codes . we have considered three cases : the standard bp decoder adapted from classical decoding to decode quantum codes , the superior bp decoder with poulin and chung s random perturbation @xcite , and the bp decoder with our new feedback adjustment strategy introduced here .
we used the @xmath8 representation of quantum codes to construct our feedback - based algorithm and exploit not just the syndrome but all the measurement resulting from stabilizer measurements .
then we used the block error rate ( ber ) and average number of iterations ( anoi ) to demonstrate a dramatic better error correcting capacity improvement with relative less iterations .
this result was achieved using new feedback adjustment strategy vs the two alternatives : standard bp decoder and the bp decoder with random perturbation .
as shown in section v , the net rate for ea-2 and ea-3 are 0 and @xmath242 respectively .
as pointed out by brun , devetak and hsieh , in general , net rates for ea quantum codes can be positive , negative , or zero .
the zero rate does not mean that no qubits are transmitted by this code !
rather , it implies that a number of bits of entanglement is needed that is equal to the number of bits transmitted . "
compared to quantum codes with high net rates , ea quantum codes with zero or negative net rates employ many more physical qubits including a great number of bits of entanglement that being assumed to be error - free , thus these codes can tolerate stronger noise .
therefore , it is not surprising that ea-2 and ea-3 greatly increase @xmath4 the cross - over probability of depolarizing channel when compared to quantum codes with similar code length and higher net rates . for comparison , based on construction b " @xcite ( same as the construction of conventional ) , we construct a quantum code with a similar code length as ea-2 but having a relative higher net rate .
we name this example as ex which has a code length of 800 and a net rate @xmath243 . in fig .
[ fig : ex ] , we see that ex has comparable ber performance to other examples presented in existing literatures which have similar parameters as ex but significantly worse than ea-2 . for the standard bp decoder and for our algorithm when applied to ex .
the maximum number of iterations is @xmath189 .
both the number of iterations between each perturbation and the maximal entries traversed by our decoder are @xmath190.,title="fig:",width=283 ] +
the authors appreciate critical comments by d. poulin on an early version of the manuscript .
yjw appreciates financial support by the china scholarship council .
bcs received financial support from _
_ i__core , and bmb and xmw received financial support from the 973 project of china under grant no .
2010cb328300 , the nsfc - guangdong jointly funded project of china under grant no .
u0635003 , and the 111 program of china under grant no .
bcs is supported by a cifar fellowship
. 1 r. g. gallager , `` low density parity - check codes , '' ire trans .
theory , vol .
it-8 , pp .
21 - 28 , jan .
d. j. c. mackay , `` good error - correcting codes based on very sparse matrices , '' _ ieee trans .
inform . theory _
399 - 431 , march 1999 .
w. shor , `` scheme for reducing decoherence in quantum computer memory , '' _ phys . rev .
a _ , vol .
r2493-r2496 , 1995 .
a. m. steane , `` error - correcting codes in quantum theory , '' _ phys .
_ , vol .
793 - 797 , 1996 .
d. gottesman , `` stabilizer codes and quantum error correction , '' ph.d .
thesis , california institute of technology , pasadena , ca 1997 ; d. gottesman .
( 2007 ) lecture notes : quantum error correction .
[ online ] .
available : http://www.perimeterinstitute.ca/personal/dgottesman/qecc2007 c. h. bennett , d. p. divincenzo , j. a. smolin and w. k. wootters , `` mixed state entanglement and quantum error correction , '' _ phys .
a _ , vol .
3824 - 3851 , 1996 .
e. knill and r. laflamme , `` a theory of quantum error - correcting codes , '' _ phys
a. _ , vol .
900 - 911 , 1997 .
a. calderbank , e. rains , p. shor , and n. sloane , `` quantum error correction via codes over gf(4 ) , '' _ ieee trans .
inform . theory _
1369 - 1387 , july 1998 .
d. j. c. mackay , g. j. mitchison , and p. l. mcfadden , `` sparse - graph codes for quantum error correction , '' _ ieee trans .
inform . theory _
2315 - 2330 , oct .
d. poulin and y. chung , `` on the iterative decoding of sparse quantum codes , '' _ quantum inform .
_ , vol . 8 , pp .
987 - 1000 , 2008 .
d. poulin , j .-
p tillich , and h. ollivier , `` quantum serial turbo codes , '' _ ieee trans .
inform . theory _
6 , pp . 2776 - 2798 , june 2009 . t. a. brun , i. devetak , and m .- h .
hsieh , `` correcting quantum errors with entanglement , '' _ science _
. 314 , pp . 436 - 439 , october 2006 . m .- h . hsieh , t. a. brun , and i. devetak , `` entanglement - assisted quantum quasi - cyclic low - density parity - check codes , '' _ phys .
a _ , vol .
79 , 032340 , 2009 .
wang , b .-
bai , w .- b .
zhao , and x .- m . wang , `` entanglement - assisted quantum error - correcting codes constructed from irregular repeat accumulate codes '' _ int .
j. quantum inf .
_ , vol . 7 , no . 7 , pp . 1373 - 1389 , 2009 . j. preskill , _ lecture notes for physics 219 : quantum computation ( 2001)_. [ online ] .
available : www.theory.caltech.edu/people/preskill/ph219 .
t. k. moon , _ error correction coding : mathematical methods and algorithms .
_ new jersey , u.s.a .
: john wiley @xmath244 sons , inc . , 2005 .
t. camara , h. ollivier , and j .-
tillich , `` a class of quantum ldpc codes : construction and performances under iterative decoding , '' in _ proc .
inf . theory _
, nice , france , june 2007 , pp .
811 - 815 . m. s. leifer and d. poulin , `` quantum graphical models and belief propagation , '' _ annals of physics _ , vol .
. 1899 - 1946 , 2008 .
a. r. calderbank and p. w. shor , `` good quantum error - correcting codes exist , '' _ phys .
a _ , vol .
pp . 1098 - 1105 , 1996 .
a. shokrollahi , `` ldpc codes : an introduction , '' tech . rep .
fremont , ca : digital fountain , inc . , apr . 2 , 2003 .
d. j. c. mackay , `` good error correcting codes based on very sparse matrices , '' _ ieee trans .
inform . theory _
2 , pp . 399 - 431 , mar .
this regular ldpc check matrix is available at : http:// www.inference.phy.cam.ac.uk/mackay/codes/en/c/816.3.174 .
this regular ldpc check matrix is available at : http:// www.inference.phy.cam.ac.uk/mackay/codes/en/c/1920.1280.3.303.gz .
r. laflamme , c. miquel , j .-
paz , and w. h. zurek , `` perfect quantum error correcting code , '' _ phys .
_ , vol .
77 , pp .
198 - 201 , 1996 .
f. r. kschischang , b. j. frey , and h .- a .
loeliger , `` factor graphs and the sum - product algorithm , '' _ ieee trans .
inform . theory _ , vol .
498 - 519 , feb . 2001 .
y. kou , s. lin , and m. fossorier , `` low - density parity - check codes based on finite geometries : a rediscovery and new results , '' _ ieee trans .
inform . theory _
47 , no . 7 , pp . 2711 - 2736 , nov .
m. hagiwara and h , imai , `` quantum quasi - cycli ldpc codes '' in _ proc .
inform . theory _
, nice , france , june 2007 , pp .
806 - 810 .
m .- h hsieh , w .- t yen , and l - y hsu ,
`` high performance entanglement - assisted quantum ldpc codes need little entanglement , '' _ ieee trans .
inform . theory _
1761 - 1769 , mar . | decoding sparse quantum codes can be accomplished by syndrome - based decoding using a belief propagation ( bp ) algorithm .
we significantly improve this decoding scheme by developing a new feedback adjustment strategy for the standard bp algorithm . in our feedback procedure ,
we exploit much of the information from stabilizers , not just the syndrome but also the values of the frustrated checks on individual qubits of the code and the channel model .
furthermore we show that our decoding algorithm is superior to belief propagation algorithms using only the syndrome in the feedback procedure for all cases of the depolarizing channel .
our algorithm does not increase the measurement overhead compared to the previous method , as the extra information comes for free from the requisite stabilizer measurements .
shell : bare demo of ieeetran.cls for journals sparse quantum codes , quantum error correction , quantum channels , belief propagation , stabilizers . |
40004000 = 1000 # 1
40004000 = 1000 brown , bildsten & rutledge ( * ? ? ?
* bbr98 hereafter ) argued that the core of a transiently accreting neutron star ( ns ) ( for reviews of transient neutron stars , see ) is heated by nuclear reactions deep in the crust during the accretion outbursts .
the core is heated to a steady state in @xmath1 yr ( see for a detailed calculation ) , after which the ns emits a quiescent thermal luminosity ( bbr98 ) @xmath3 where @xmath4 is the time - averaged ( over the ns core thermal timescale ) mass - accretion rate onto the ns , and @xmath5 is the amount of heat deposited in the crust per accreted nucleon (; see for a review ) . for an accretion flux onto the ns ( @xmath6 ) ,
the `` rock bottom '' quiescent flux due to deep crustal heating is then @xmath7 where @xmath8 is the accretion flux averaged over the ns core thermal timescale , and the accretion efficiency @xmath9 for accretion luminosity of @xmath10 .
since @xmath11 , where @xmath12 is the mean outburst duration and @xmath13 is the mean recurrence timescale , this scenario relates the quiescent luminosity to the outburst properties ; comparisons to observations of several quiescent neutron stars ( qnss ) match the predictions reasonably well @xcite .
it also helps us to understand why the black hole systems in quiescence are so much fainter than ns systems @xcite , as they can not be thermally emitting .
equation ( [ eq : fq ] ) is the minimum quiescent flux from the neutron star .
should residual accretion occur in quiescence @xcite , the accretion luminosity would be in addition to the thermal emission already present .
is a transient , type - i x - ray bursting neutron star approximately 3.8 degrees from the galactic center @xcite .
it was discovered in outburst in 1989 , when its luminosity was 1.3 ( d/8 ) @xmath14 ; subsequent analysis found that it had been in outburst at least since oct 1988 ( ) , and has remained bright since then ( see for a complete description of observations ) .
sunyaev et al .
( 1990 ) suggested that it was a transient , apparently because it had not previously been detected , although no strong upper limits on previous x - ray emission were given .
( this was noted later by , who suggested the object may not be a transient at all .
we adopt an outburst timescale of @xmath15 13 yr , although the outburst may have been longer ) .
type i x - ray bursts were observed , at the rate of @xmath110 per day , and studied in detail with rxte @xcite .
nearly coherent oscillations ( 580 hz ) have been observed during the type - i bursts , and the frequency is interpreted as the spin frequency of the ns @xcite .
the uncertainty in the localization of this source was initially 4.2 ( 90% confidence ) . an improved error circle ( 10 with _ rosat_/hri ; ) found 13 candidate ir counterparts down to @xmath16 ( 10 @xmath17 ) .
the integrated hi column density in the direction of is = 0.35 ( @xmath18 cm )
. found = 16 ( @xmath19=534 ; ) , suggesting that is variable , and thus there must be significant contribution to from the system itself .
an _ rxte_/pca scan of the galactic center found that had entered a low luminosity state @xcite in early 2001 . then observed the source with _ chandra_/acis - s in late march ( @xmath11 months after the end of the outburst ) and detected it at an unabsorbed flux of @xmath20 ( 0.510 kev ) .
the scenario described by equation ( [ eq : fq ] ) applies when the neutron star crust and atmosphere resemble that of a cooling neutron star that is , the temperature decreases with increasing radius . for ns transients with short outbursts , such as aql x-1 ( @xmath21 ) ,
this is a good approximation , as the increase in crust s temperature from the heating during the outburst is small @xcite .
this is not the case for .
the duration of its most recent outburst is of the order of ( or longer than ) the thermal diffusion timescale in the crust ( bbr98 ) . in a sense , in quiescence resembles a neutron star that accreted steadily at the _ outburst _ accretion rate , except that the core temperature will be at the lower value set by the time - averaged accretion rate ( over the previous @xmath22 ) .
to illustrate how the crust is so dramatically heated during a long outburst , consider the rise in temperature if no heat were conducted away from the crust during the outburst . as we describe in
[ sec : crust ] , the total heat capacity of the region of the crust where most of the heat is deposited is @xmath23 erg k@xmath24 . during the outburst ,
the total amount of heat deposited is @xmath25 and so , _ if no heat is conducted away from the heating region _ , the temperature there can rise to @xmath26 k during the 13 year outburst . even when thermal conduction is taken into account , the rise in crust temperature can still be @xmath27 , which is the typical temperature the crust and core would have if accreted steadily at @xmath28 yr@xmath24 ( for @xmath29 yr ; see
[ sec : core ] ) .
if the crust , which is composed of the ashes of h / he burning , has a low thermal conductivity , then there is a substantial temperature gradient between crust and core . as discussed by for the case of steadily accreting neutron stars , when there is a substantial thermal gradient in the inner crust , the temperature of the crust reaction layers becomes decoupled from that of the core . for the ns in ,
this means that , until the crust has thermally relaxed , we can not directly infer its core temperature , as is possible for short - outburst transients such as .
however , this also means that one can directly compute the current quiescent luminosity from using the fluence during the last outburst , as the thermal state of the crust is only weakly sensitive to the temperature of the core , and , hence , is nearly independent of the uncertainty in the accretion history over the past @xmath30 yrs ( the thermal time of the core ) , unlike in the case of short - outburst transients .
our simulations , presented here , predict a quiescent luminosity which agrees with the observed value to within observational and theoretical uncertainties ( factor of a few ) for the fiducial value of @xmath5 and outburst fluence .
moreover , this dependence of the thermal flux in the crust - dominated regime on @xmath31 , and not on @xmath32 , allows us to predict the evolution of @xmath33 , which can be directly confronted with observations .
the domination of the quiescent luminosity by the cooling crust sets apart from the other neutron star transients with short duration outbursts and makes it an ideal laboratory for separately measuring the thermal properties of the crust .
in this paper we report our analysis of the _ chandra _ observation @xcite of in quiescence and describe our simulations of the thermal state of its crust and core .
we begin in [ sec : anal ] by describing the _ chandra _ observation and spectral analysis of .
we include a description of possible ir counterparts ( [ sec : ir ] ) . after describing the observations
, we then discuss the implications for the crust and core thermal structure as outlined above .
we first apply ( [ sec : core ] ) the scenario in which the crust heating for any single outburst is negligible , i.e. , we treat as if it were a short - duration outburst , such as aql x-1 .
this analysis has been done previously by .
we then explain , in [ sec : crust ] , why this analysis is inapplicable to this source , and how the quiescent luminosity is determined by the thermal properties of the crust ( its thermal diffusion timescale and heat capacity ) for the next 130 yr . we present simulations of the quiescent lightcurve and demonstrate that monitoring observations can constrain these properties , similar to proposals for glitching pulsars @xcite .
we conclude in [ sec : con ] with a summary and discussion of these results .
the data were obtained from the _ chandra _ public archive .
the discovery and observations have been previously analyzed by , and details of the trigger and history of are included there .
the observation was triggered when an _
rxte_/pca scan of the galactic center found that had entered a low luminosity state @xcite , and was made 2001 mar 27 00:18 - 06:23 ( tt ) with the acis - s instrument @xcite for a total exposure time of 19401 seconds .
the x - ray source was imaged on the s3 ( backside illuminated ) chip , which was operated as a 1/4 array with 0.8 sec exposures .
we analyzed the data using the ciao v2.1 with caldb v2.6 and xspec v11 @xcite .
two x - ray point sources separated by 31.30.1 are found with _ celldetect _ above a signal / noise ratio of 5 , listed in table [ tab : objs ] .
we compared this image with an archived _ rosat_/hri image of this region ( rh400718 ) taken in 1997 when was in outburst .
the position of had been previously determined this way @xcite and clearly is at the position of _ chandra _ source # 1 .
x - ray source counts were extracted within an area 5 pixels in radius about the source position , for a total of 183 counts . at 0.0075 counts
/ frame , the pileup fraction is negligible .
background was taken from an annulus centered on the source position , with radii of 10 and 80 pixels .
the expected number of background counts in the source region is 3 .
a ks test @xcite finds that the times of arrival ( toa ) of the 183 counts is consistent with a constant countrate .
we place 3@xmath17 upper - limits on variability of @xmath34=22% rms .
we also produced a power - density spectrum to search for a pulsed signal ( number of frequency bins : 12750 , with frequency resolution of 5 hz , and a nyquist frequency of 0.625 hz ; see ) , using barycentered toas ( tool _ axbary _ ) .
no evidence for a pulsed signal is found : the largest leahy - normalized power @xcite was 20.37 ( with a probability of chance occurrence from a poisson - distributed countrate of 0.48 ) .
the absence of a coherent signal is not surprising in light of the detection of pulsations during type - i x - ray bursts @xcite at @xmath1580 hz , well above our nyquist frequency .
we binned the data into 10 pi bins ( 0.5 - 10.0 kev ) , and fit several single component spectral models ( powerlaw , h atmosphere , or raymond - smith , a multicolor disk and blackbody ) .
galactic absorption is initially left as a free parameter .
the best - fit models were all statistically acceptable ; the parameters are given in table [ tab : fits ] . while all models we investigated are statistically acceptable , some can be argued against on physical grounds .
the pure power - law spectrum is unusually steep ( @xmath35=5.20.6 ) , typical more of a thermal spectrum than other processes .
the emission measure @xmath36 from the raymond - smith spectrum is higher by 2 orders of magnitude than typical from active stellar coronae in the analogous rs cvn systems (;
see for discussion on coronal emission from companions in x - ray transients ) .
the disk black - body spectral model implies an inner disk radius of @xmath10.7 km considerably smaller than a neutron star .
our best - fit blackbody spectrum is consistent with that found by . as pointed out by bbr98 , for accretion rates 2 @xmath37yr , gravity stratifies metals in the ns atmosphere faster than they can be provided by accretion ( bildsten , salpeter , & wasserman ) , making a pure h atmosphere the appropriate description of the ns photosphere in quiescence . due to the strongly energy - dependent opacity of free - free transitions , these spectra are significantly different from black - bodies @xcite . unlike the results from black - body fits , the inferred ns radii ( = @xmath38 where @xmath39 is the proper radius , and @xmath40 means the value as measured by a distant observer ) of the four observed field transients using h atmosphere spectra are consistent with what are expected theoretically : @xmath112 km ( 4 , ; , ;
1608 - 522 ; and 4u 2129 + 47 ) , thus supporting the notion that much of the emission originates from a neutron star surface .
the quiescent spectrum for qnss is thus interpreted as thermal emission from a pure hydrogen atmosphere ns , possibly with an underlying power - law ( whose origin is not understood ) which dominates the spectrum at high ( @xmath413 kev ) energies ( see rutledge ) .
the spectral fit for a h atmosphere spectrum alone ( no power - law ) gives @xmath42=12030 ev , an emission area radius of = 10 ( d/8 kpc ) km , and an unabsorbed flux of 1.8 ( 0.5 - 10 kev ; the absorbed flux is 3.8 ) , corresponding to a luminosity of @xmath431.4 ( d/ 8)@xmath14 ( 0.5 - 10 kev ) .
the 90% confidence range in the unabsorbed flux is ( 1.14.6 ) . the best - fit spectrum is shown in fig .
[ fig : ks1731 ] . the bolometric thermal luminosity ( as observed at infinity ) is @xmath442.7 ( d/8 ) @xmath14 ( uncertain by a factor of @xmath13 , due largely to spectral uncertainty ) . if 100% of the emergent luminosity were from accretion ( at an efficiency of 0.2 ) , this sets @xmath45 @xmath37yr , sufficiently low for the assumption of an h atmosphere .
when we include an underlying power - law with all parameters free , the s / n of the spectrum does not constrain the model parameters to better than an order of magnitude .
when we hold and @xmath35 fixed at typical values ( = 12.5 km for d=8 kpc ; @xmath46 ; cf . ) , the spectral model provides an acceptable fit to the data .
this demonstrates that the spectrum of in quiescence is consistent with that observed from other qnss .
a comparison between chandra astrometry and 2mass images finds a possible ir counterpart at the position of source # 2 ( see below ) but not of , with limits in j , h and k@xmath47 of 15.8 , 15.1 and 14.3 magnitudes respectively .
comparing with ir sources previously examined as possible counterparts to @xcite , the closest object is source `` h '' , which is 1.51.1 away ( we adopt 1 positional uncertainty for _ chandra _ , and 0.5 uncertainty for the ir source ) . the second closest is `` g '' , which is 3.3 away .
in addition , there is a fainter ir object present in the h - band image ( `` new '' object ) , but not in the j - band image taken by barret , which is not marked , but which is consistent with the _ chandra _ position for .
further work in the ir for example , searching for ellipsoidal variations in quiescence is needed for a positive identification of the counterpart .
source # 2 is consistent in position with 2mass j173412.7 - 260548 ( @xmath48 in distance , about the accuracy of the 2mass position reconstruction error ) . for an ir field source density of @xmath49=6.9 arcsec@xmath50 ( 217700 objects in the 2mass catalog within a 1000 radius )
the probability of a chance coincidence is prob=@xmath51=3.9% , which is the significance of the association of the x - ray and ir objects .
the source s 2mass magnitudes and colors are j=12.2640.033 , @xmath52=0.948 and @xmath53=1.33 .
[ [ core - dominated - emission - crust - dominated - emission - and - the - future - lightcurve - of ] ] core dominated emission , crust dominated emission , and the future lightcurve of ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ having described the observations and spectral analysis of , we now turn our attention to the thermal properties of the neutron star s crust and core .
we begin by estimating the amount of mass deposited during the previous outburst ; we then describe what the thermal structure would be if the change in crust temperature during this outburst were small ( bbr98 ; ) .
this is the case for transients with short - duration outbursts .
because of the length of the outburst for this source , however , the crust is substantially heated ; [ sec : crust ] describes how the observed emission is determined by the heating of the crust during the past outburst , and not by the core equilibrium temperature . using the method of ur01
, we simulate the evolution of the quiescent lightcurve for different regimes of the crust and core microphysics .
has been observed several times , always in outburst , since 1988 , and nearly continuously so since jan 1996 with the _ rxte_/asm . to estimate @xmath54 , we integrated the asm counts using only data in which there was a @xmath55 1-day detection and taking only 1 day detections in which the previous observation was made @xmath564 days ( to insure adequate sampling ) prior .
periods in which the time to the previous observation was @xmath414 days were not included .
the total fluence is 13660 asm c / s@xmath57days , over an integration of 1599 days ( including both detections and non - detections ) .
if we assume an additional countrate at the 2@xmath17 level on those observations when the countrate was below 3@xmath17 detection , the total fluence would only increase by 7% , which is our 2@xmath17 upper - limit for the uncertainty in integrated counts . assuming a @xmath58=5 kev bremsstrahlung spectrum with = 1.0 ( which gives 6.5 per 1 asm c / s , corrected for absorption , 0.0120 kev , from w3pimms ; in the 210 kev range , the conversion factor is 3.0 per 1 asm c / s )
, we find a mean outburst flux , corrected for absorption , of @xmath59 ( 6.5 per 1 asm count ) = 5.6 .
we conservatively estimate the spectral uncertainty to be at the 40% level ( for a change in @xmath58 between 2 kev and 8 kev , the asm counts / flux conversion changes by 40% ) . extrapolating this mean to the entire outburst
, we find an outburst fluence of @xmath60 . this estimate is greater by a factor of @xmath16 than that of , who estimated @xmath61 using the minimum observed outburst flux and an outburst duration of 11.5 yr , rather than 13 yr .
the quiescent bolometric thermal flux is @xmath62 .
if the change in crust temperature during the outburst were small , then equation ( [ eq : fq ] ) would apply : the time - averaged accretion flux is @xmath63 , so that @xmath64 there is an uncertainty of about 0.5 dex in this value , mostly due to the uncertainty in the bolometric value of @xmath65 .
this estimate is similar to the one arrived at independently by , although our estimated mean recurrence time is longer ( 1500 0.5 dex yr vs. 200 yr ) , since our time - averaged outburst flux is a factor of six greater than that assumed by . strictly speaking ,
this estimate of @xmath13 is the recurrence time to an outburst of comparable fluence .
a month - long outburst would have just 1% of the fluence of the previous outburst and would not affect the estimate of eq .
( [ eq : rec ] ) .
this estimate also neglects any neutrino emission , which reduces the effective value of @xmath5 and hence @xmath13 .
more seriously , however , this analysis ignores the fact that the crust is strongly perturbed away from the core equilibrium temperature during such a long outburst . as a result
, the quiescent emission currently observed is set by the thermal relaxation of the heated crust , and as we now demonstrate , is mostly decoupled from the thermal state of the core . during an outburst , the heating from the reactions raises the temperature in the crust around the reaction layers .
after the outburst ends , the crust thermally relaxes and the thermal profile comes to resemble a cooling neutron star , i.e. , the temperature decreases with radius . for neutron star transients with `` typical '' outburst fluences and recurrence times ( that is , @xmath66 month , with an outburst luminosity @xmath67 eddington ) , the variations in the crust temperature and hence @xmath68 are small ( bbr98 ; ur01 ) . for example , ur01 found that @xmath68 varied by @xmath1 few% for @xmath69 yr and by @xmath1 30% for @xmath70
the variation is small because the amount of energy deposited in such short outbursts is not significant compared to the heat content of the crust , and so the crust is not heated substantially compared to its temperature when in thermal equilibrium with the core . to ascertain the magnitude of the deviation of @xmath33 from the value predicted by eq .
( [ eq : brown ] ) , we performed several simulations of thermal relaxation of the ns crust in a transient with observational properties ( @xmath12 , @xmath13 , @xmath72 ) inferred for using the methods and microphysics as described in ur01 . in summary , we solve the non - relativistic heat equation in the crust , from @xmath73 g @xmath74 to @xmath75 g @xmath74 .
heating due to nuclear reactions in the deep crust is simulated by depositing energy at densities corresponding to the nuclear transitions computed by , with the amount of energy set by the instantaneous accretion rate . at the outer boundary ,
we use the flux - temperature relation for a fully accreted crust from potekhin @xcite .
the core is taken to be isothermal ( a good assumption for the timescales of interest because of the large thermal conductivity ) , and its temperature evolves according to the mismatch between its neutrino emissivity and the flux from the crust .
we start our simulations with the temperature profile corresponding to persistent accretion at the rate @xmath32 corresponding to the time average of @xmath31 over the recurrence time .
we then evolve the model through several outburst / quiescence cycles , as necessary for the model to `` forget '' the initial conditions and reach a limit cycle . if the crust resembles that of a neutron star steadily accreting at @xmath31 , then the estimate of the recurrence time ( eq . [ [ eq : rec ] ] ) is inapplicable ( except as a lower limit ) .
this is because the crust temperature is decoupled from that of the core .
a full survey of parameter space should include simulations over a range of @xmath13 .
this is beyond the scope of this initial paper ; rather , we presume that @xmath13 is given by eq .
( [ eq : rec ] ) in order to survey the influence of the crust and core microphysics . in these simulations
, we examine the two main uncertainties in ns microphysics , ( 1 ) the impurity fraction , and , hence , the conductivity of the crust and ( 2 ) the possible presence of `` enhanced '' core neutrino cooling mechanisms , such as direct urca or pion condensation .
first , the composition of accreting ns crusts is set by the nuclear processing of products of burning in the upper atmosphere .
published calculations @xcite presume that burning proceeds to pure iron , and , hence , the crust has no impurities .
however , recent work by schatz and collaborators @xcite shows that burning products are a mix of elements substantially heavier than iron , with the impurity parameter of the elements of the mix from the average charge @xmath76 , @xmath77 , where the average takes into account the relative abundance of the species . ]
@xmath78 comparable to the square of the average nuclear charge @xmath79 .
we therefore set bounds on the crustal conductivity by using electron - ion scattering @xcite , which has the same form as electron - impurity scattering with @xmath80 , and electron - phonon scattering @xcite , which is appropriate for a pure crystal .
we refer to these two cases as `` low @xmath81 '' and `` high @xmath81 '' , respectively .
second , the uncertainty in core cooling is covered by simulations with `` standard cooling '' and `` enhanced cooling '' ( modified urca appropriately suppressed by nucleon superfluidity , and neutrino emission as when the core is a pion condensate , respectively ; see ur01 for details ) . to illustrate the deviation of the crustal temperature from the equilibrium core temperature and the subsequent thermal relaxation of the crust
, we show in fig .
[ fig : tevol ] the time evolution of the crustal temperature in the low @xmath81 , standard cooling case for observational parameters inferred for ( @xmath82 yr , @xmath83 yr , and @xmath84 yr@xmath24 ) . vertical slices of the surface in the density - temperature plane show the instantaneous temperature distribution in the crust , while slices in the time - temperature plane show the time evolution of the temperature since the beginning of the outburst . clearly , the crust is heated to well above its pre - outburst temperature profile , which tracks the core temperature ( note that , because the crust is heated to well above its temperature in equilibrium with the core , this qualitative result is not very sensitive to the assumed @xmath13 ) .
a steep temperature gradient carries most of the heat from the deep crustal heating region around @xmath85 g @xmath74 into the core , while a shallower temperature gradient carries a fraction of the heat towards the surface .
the temperature near the top of the crust ( @xmath1 g cm ; and , hence , the x - ray luminosity , @xmath86 ; ) reaches a maximum at the end of the outburst ( time=13 years ) , and then decays back to the pre - outburst value on the thermal timescale of the crust ( @xmath87 yrs for this model ) .
thermal evolution in other cases is similar ( but , of course , is different in the absolute magnitude of the temperature change ) , and the temperature changes are always much bigger than those that occur during short - duration outbursts ( cf.fig . 1 of ur01 ) . in fig .
[ fig : greg ] , we show the long - term evolution of the quiescent surface luminosity from after accretion onto the compact object has ceased for the four cases discussed above , namely , low and high crustal conductivity @xmath81 , and standard and enhanced core neutrino emission ( see the legend on the plot ) .
first consider fig .
[ fig : greg]a , where we set @xmath88 .
in all four cases , the transition from crust - dominated ( at early times ) to core - dominated cooling ( at late times ) is evident as a drop in the luminosity , ranging from 50% ( high @xmath81 , standard cooling case ) to a factor of @xmath89 ( low @xmath81 , enhanced cooling case ) .
regardless of the core neutrino emissivity , this transition occurs at @xmath87 yrs for low conductivity crusts , and after @xmath90 yr for high conductivity crusts .
these timescales are easy to understand , as they are just the thermal time to the appropriate depth in the crust ( cf . figure 3 of ur01 ) .
since the low @xmath81 crust has a longer thermal time , it stays hot about 30 times longer than the high @xmath81 crust .
after the crust thermally relaxes , @xmath33 is set by the emission from a hot core .
if the only core neutrino emission mechanism is modified urca , suppressed by nucleon superfluidity , then the amount of heat lost by neutrinos from the core is negligible .
therefore , all the heat deposited into the star comes out as thermal emission , and @xmath33 asymptotically approaches the value given eq .
( [ eq : brown ] ) , with @xmath91 ( see solid and dashed lines in fig . [
fig : greg ] ) . on the other hand , when enhanced neutrino cooling is allowed , @xmath92% of the deposited heat escapes as neutrinos (;
ur01 ) , and only the remaining @xmath93% is radiated thermally from the surface . in this case , the `` effective '' value of @xmath5 in eq .
( [ eq : brown ] ) , in the sense of the energy retained in the star and re - radiated thermally , is @xmath94 .
as shown above , the crust is heated to well above the core temperature during the long outburst .
how does the temperature difference between the crust and the core depend on the crustal conductivity ? as an estimate
, we can assume that , during the outburst , the crust comes to an equilibrium where the heat deposited around @xmath95 g @xmath74 at a rate @xmath96 is conducted away from this region .
this is a very good assumption for the high conductivity case , and an acceptable one for the low conductivity case , where the crustal thermal time is comparable to the outburst duration . during the outburst , most of the heat is conducted into the core with a flux @xmath97 , where @xmath98 m is the distance between the crustal heating region and the core .
thermal balance then requires @xmath99 , or @xmath100 k @xmath101 , where @xmath102 is the conductivity in units of @xmath103 erg cm@xmath24 s@xmath24k . in the low conductivity case , @xmath105 , so @xmath106 .
we see from fig .
[ fig : greg]a that @xmath33 at early times in the low @xmath81 case is the same for standard and enhanced cooling ( solid and dotted lines ) .
this is now easy to understand , since , as we showed above , the crustal temperature is decoupled from the core temperature if the conductivity is low .
this situation is exactly the same as for persistent accreters discussed by brown @xcite . on the other hand , for high conductivity crusts , @xmath107 , so @xmath108 can not significantly deviate from @xmath109 during the outburst .
therefore , crust - dominated @xmath33 in the high @xmath81 case ( dashed and dash - dotted lines in fig .
[ fig : greg]a ) more closely reflects the core temperature , which is quite different between the standard and enhanced cooling cases . from examination of lightcurves in figs .
[ fig : greg]a , it is likely that was observed in quiescence during the crust - dominated phase , and is presently evolving through this phase . in three of four cases considered in fig .
[ fig : greg]a , the predicted quiescent luminosity @xmath33 right after the outburst is a factor of 3 higher than the observed @xmath110 ( itself uncertain by a factor of 3 , due to spectral uncertainty ) .
since @xmath33 depends directly on @xmath111 over the outburst duration , this difference could also be due to a systematic overestimate of @xmath31 during the @xmath112 yr outburst , only the last 5 years of which have been covered @xmath1 daily with rxte / asm .
alternatively , this difference could be due to a value of @xmath5 different from the fiducial , since the exact value of @xmath5 is uncertain ( see schatz 1999 , 2001 for a discussion on different crustal compositions than assumed by ) . to provide lightcurves which extrapolate from the currently observed luminosity , in fig .
[ fig : greg]b we adjusted @xmath5 ( or , equivalently , @xmath113 ) by factors of 0.3 to 2 , as noted in the figure caption , to obtain @xmath114 erg s@xmath24 at the start of quiescence .
we stress that the overall normalization of these lightcurves is uncertain due to uncertainty in @xmath5 and outburst fluence , but the shape is determined by the microphysics of the crust and the temperature of the core .
the time evolution of the luminosity will permit distinction between the four cases , and , hence , directly infer the integrated conductivity of the crust and possibly indicate whether enhanced neutrino emission mechanisms are operating in the core .
in particular , observing the drop in the luminosity will permit the measurement of the thermal timescale of the crust .
in addition , the relative magnitude of the drop will tell us about the presence or absence of enhanced neutrino emission from the core .
such measurements ( accurate to 10% ) are well within the capabilities of present x - ray instrumentation , as the present detection demonstrates .
observations over the next year will be able to exclude the case where the crust is a pure crystal and hence has very high conductivity .
since the initial understanding of type i bursts as thermonuclear flashes , there has been an open question as to how much flux is rising up from deep parts of the star into the burning region ( see ) .
large values of this flux ( @xmath115 mev per accreted nucleon ) can stabilize burning at accretion rates lower than otherwise expected .
first showed that most of the crustal heating in a persistent source went into the ns core and that , at most , 100 kev per accreted nucleon came out into the upper parts of the star .
the quiescent observations of shortly after outburst strongly support this picture and tell us that deep nuclear heating will not impact the type i x - ray burst behavior . assuming that the source has been transiently accreting for the past @xmath116 or so , with outburst fluences similar to that of the most recent observed outburst
, we can constrain the core temperature to be less than the maximum crust temperature . to determine the most conservative (
i.e. , largest ) upper bound on this peak crust temperature , we use the following assumptions .
we take the atmosphere to be composed of pure iron ( which has a higher opacity than hydrogen / helium ) and we use low thermal conductivity ( electron - ion scattering ) in the crust .
these assumptions lead to the largest temperature rise between the photosphere and the crust .
taking @xmath117 , we find a peak crust temperature , and hence a maximum core temperature , of @xmath118 .
the large amount of heat deposited during the recent ( and the only observed ) outburst of , and the lengthy duration of the outburst compared to the thermal diffusion time of the ns crust , imply that the quiescent flux recently observed is dominated by emission from the cooling crust and not the core .
the importance of this result is that the quiescent luminosity _ is calculated using the properties of only the most recent outburst _ , rather than by estimating the accretion history over the past yr , as is the case in the core - dominated transients such as and 4@xcite .
our prediction of the quiescent luminosity from the cooling crust following the outburst agrees , to within observational and theoretical uncertainties ( a factor of a few ) , with the estimate of the observed bolometric luminosity . unlike the case of the short @xmath12 transients where we can directly infer the average core temperature , in the case of we can not measure the core temperature , and hence the recurrence time , until after the crust has thermally relaxed ( 130 yr after the end of the outburst ) . assuming that the source has been transiently accreting for the past @xmath116 or so , with outburst fluences comparable to that of the most recent outburst
, we can constrain the core temperature to be less than the peak crust temperature , which , using the relation of for an iron crust ( low conductivity ) , gives @xmath119 .
we predict future lightcurves for on the basis of 4 different scenario for crustal and core microphysics ( high / low conductivity , and enhanced / normal neutrino emissivity ) .
these can be distinguished from each other observationally one in @xmath561 yr ( excluding or confirming the high @xmath81 crust with enhanced neutrino emissivity core ) , others over longer terms ( 10 - 100 yrs ) .
this offers the unique opportunity to directly constrain the properties of the ns crust thermal conductivity , heat capacity , and depth to the ns core through monitoring observations that would measure the predicted decaying lightcurve .
is now the fifth after 4 , , 4u
1608 - 522 , and 4u 2129 + 47field transient neutron star which displays in quiescence a spectrum consistent with a h atmosphere ns of radius @xmath1 12 km , and ( possibly ) a hard ( @xmath1200.85 ) power - law component which comprises 1040% of the quiescent ( 0.510 kev ) flux . shows no variability on short ( @xmath56hours ) timescales .
these spectral and variability properties are identical to those of the well - studied and 4 @xcite , and the spectral properties are similar to those from lower s / n data of 4u 1608 - 522 and 4u 2129 + 47 @xcite .
the derived radius is also comparable to that derived from this object using the independent method of a cooling radius expansion burst @xcite .
it is possible that a persistent source can turn off after long periods of accretion ( here , a persistent source is one accreting for a period @xmath121 yr , the core heating timescale ) . while eq .
1 would seem to be applicable to predict the post turn - off quiescent thermal luminosity from the crust , the equation assumes that all heat deposited in the crust is re - radiated during quiescence
. that will not be the case for sources approaching eddington luminosity , such as , where the effect of neutrino cooling in the crust and core can not be neglected .
in such cases , @xmath122 . here @xmath123 is the energy per accreted baryon emitted through the atmosphere from the crust immediately following the outburst .
a minimum value , @xmath124 ev / nucleon , applies when the source accretes persistently at the eddington limit , in which only 10% of the deposited heat flows towards the surface @xcite . at lower persistent luminosities , however , the crust and core temperatures are lower and neutrino cooling in each becomes less important , so @xmath123 increases . in the case of ( @xmath125 ) , we obtain @xmath126250 ev , if were a persistent accreter .
however , the quiescent luminosity of gives instead @xmath127 ev , a factor of 7 below the lower limit , even for a persistent source accreting at the eddington rate .
this demonstrates that can not be a persistent accreter , because its crust ( and by extension the core ) is too cool to have been accreting for @xmath121 yr .
note , however , that we have neglected the effect of of neutrino emission due to cooper pairing of superfluid particles ( see for a review ) which may reduce the value of @xmath123 .
as discussed by , the study of qnss in globular clusters has yielded a new way to accurately measure ns radii .
the distances to some globular clusters have been determined to 2% post - hipparcos @xcite , and can be measured even more directly with sim @xcite , effectively removing distance as an uncertainty . at present , there is only one globular cluster qns in which the thermal component has been spectroscopically analyzed ( in @xmath128 cen ; ) and there are two more proposed ( in 47 tuc ; ) .
measuring ns radii to an accuracy of 0.5 km even without knowing the mass can exclude @xmath150% of the proposed equations of state for the ns core matter @xcite . even though we presently can not make such accurate radius measurements in the field transients due to uncertain distances ( which can be overcome with sim ) , they remain important targets for detailed studies of the observational phenomena of qnss , such as we report here .
the relative luminosity of the ( possible ) power - law component poses a puzzle .
it has been suggested that this component is due to accretion onto the ns magnetosphere @xcite .
however , such accretion would be unrelated to the quiescent luminosity of nss as predicted from deep crustal heating ( which is dominated by outburst accretion ) , and it can only be ascribed to coincidence that in the cases where adequate sensitivity is obtained , the power - law component produces a comparable fraction of the thermal flux in different sources .
this suggests that the power - law component and h atmosphere spectral component are perhaps more closely related than previously thought .
the authors are grateful to the _ chandra _ observatory team for producing this outstanding observatory , and to the _ chandra _ data processing team who pre - handle all the data and provide the calibrations which are used in this work .
we also thank a. y. potekhin for reading and for comments on the text prior to submission .
this research has made use of data obtained through the high energy astrophysics science archive research center online service , provided by the nasa / goddard space flight center ; and of the nasa / ipac infrared science archive , which is operated by the jet propulsion laboratory , california institute of technology , under contract with the national aeronautics and space administration .
this research was partially supported by the national science foundation under grant no .
phy99 - 07949 and by nasa through grants nag 5 - 8658 , nag 5 - 7017 , nag 5 - 10865 and the _ chandra _ guest observer program through grant nas go0 - 1112b .
l. b. is a cottrell scholar of the research corporation .
e. f. b. acknowledges support from an enrico fermi fellowship .
g. u. is a lee a. dubridge fellow .
lr + & 1.0 + @xmath42(ev ) & 12030 + ( km ) & 6.5 + total model flux & 1.8 + /dof ( prob ) & 1.06/7 ( 0.39 ) + + & 1.30.3 + @xmath42(ev ) & 90 + ( km ) & 23 + @xmath35 & -0.1 + @xmath129 & 0.2 + total model flux & 3.1 + /dof ( prob ) & 0.43/5 ( 0.83 ) + + & 1.060.08 + @xmath42(ev ) & 1113 + ( km ) & ( 12.5 ) + @xmath35 & ( 0.85 ) + @xmath129 & 0.2 + total model flux & 2.1 + /dof ( prob ) & 0.47/7 ( 0.86 ) + + & 1.70.4 + @xmath35 & 5.20.6 + total model flux & 20.0 + /dof ( prob ) & 0.94/7 ( 0.48 ) + + & 0.7 + @xmath130 @xmath131 ) & ( 1 ) + @xmath132 ( kev ) & 1.40.1 + @xmath133 & ( 6.21.5 ) + total model flux & 0.93 + /dof ( prob ) & 2.32/7 ( 0.02 ) + + & 1.10.2 + @xmath134 ( ev ) & 37050 + @xmath135 ( km ) & 0.7 + total model flux & 2.0 + /dof ( prob ) & 1.1/7 ( 0.37 ) + + & 0.900.2 + @xmath42 ( ev ) & 30040 + ( km ) & 1.3 + total model flux & 1.3 + /dof ( prob ) & 1.11/7 ( 0.35 ) + | the type - i x - ray bursting low mass x - ray binary was recently detected for the first time in quiescence by wijnands , following a @xmath013 yr outburst which ended in feb 2001 .
we show that the emission area radius for a h atmosphere spectrum ( possibly with a hard power - law component that dominates the emission above 3.5 kev ) is consistent with that observed from other quiescent neutron star transients , = 23(d/8 ) km , and examine possible ir counterparts for . unlike all other known transient neutron stars ( ns ) , the duration of this recent ( and the only observed ) outburst is as long as the thermal diffusion time of the crust .
the large amount of heat deposited by reactions in the crust will have heated the crust to temperatures much higher than the equilibrium core temperature . as a result ,
the thermal luminosity currently observed from the neutron star is dominated not by the core , but by the crust .
this scenario implies that the mean outburst recurrence timescale found by wijnands ( @xmath1 200 yr ) is a lower limit .
moreover , because the thermal emission is dominated by the hot crust , the level and the time evolution of quiescent luminosity is determined mostly by the amount of heat deposited in the crust during the most recent outburst ( for which reasonable constraints on the mass accretion rate exist ) , and is only weakly sensitive to the core temperature . using estimates of the outburst mass accretion rate , our calculations of the quiescent flux immediately following the end of the outburst agree with the observed quiescent flux to within a factor of a few . in this paper
, we present simulations of the evolution of the quiescent lightcurve for different scenarios of the crust microphysics , and demonstrate that monitoring observations ( with currently flying instruments ) spanning from 130 yr can measure the crust cooling timescale and the total amount of heat stored in the crust .
these quantities have not been directly measured for any neutron star .
this makes a unique laboratory for studying the thermal properties of the crust by monitoring the luminosity over the next few years to decades .
ifundefinedchapter # 1 |
neutrino flavor transformations are determined by the elements of the pmns matrix @xcite and the differences between the squares of neutrino mass eigenvalues .
pmns matrix represents the mixture between flavor and mass eigenstates of neutrinos and is conventionally decomposed as shown in ( [ pmns ] ) .
neutrino oscillation experiments can yield the best estimations for some of those parameters , which has been demonstrated by sno @xcite and kamland @xcite .
@xmath0 \nonumber \\ & = & \left [ \begin{array}{ccc } 1 & 0 & 0 \\ 0 & c_{2 3 } & s_{2 3 } \\ 0 & -s_{2 3 } & c_{2 3 } \end{array } \right ] \left [ \begin{array}{ccc } c_{1 3 } & 0 & s_{1 3}e^{-i\delta } \\ 0 & 1 & 0 \\ -s_{1 3}e^{i\delta } & 0 & c_{1 3 } \end{array } \right ] \nonumber \\ & \times & \left [ \begin{array}{ccc } c_{1 2 } & s_{1 2 } & 0 \\ -s_{1 2 } & c_{1 2 } & 0 \\ 0 & 0 & 1 \end{array } \right ] \left [ \begin{array}{ccc } e^{i\alpha_1/2 } & 0 & 0
\\ 0 & e^{i\alpha_2/2 } & 0 \\ 0 & 0 & 1 \label{pmns } \end{array } \right],\end{aligned}\ ] ] where @xmath1 , @xmath2 , @xmath3 is the phase factor ( non - zero if neutrino oscillation violates cp symmetry ) . @xmath4 and @xmath5 majorana phase factors ( non - zero only if neutrinos are majorana particles ) , to which neutrino oscillation experiments are not sensitive . besides the pmns matrix , neutrino oscillations depend on mass eigenvalues or , more precisely , on the difference between the squared mass eigenvalues .
if there are three neutrino mass eigenvalues , then there are only two independent differences , the third being either a sum or a difference of the other two .
neutrinos studied in experiments are produced in certain flavor eigenstates with known abundances of each of them or , as an important special case , in only one flavor eigenstate .
for example , neutrinos are generated in the atmosphere with the known @xmath6 ratio of about two for low energies ; solar neutrinos and reactor antineutrinos are , initially , all @xmath7 and @xmath8 , respectively .
detector sensitivity is , generally , flavor dependent .
in particular , the inverse beta decay , the primary method for detecting reactor antineutrinos since the very beginning of neutrino experiments @xcite , involves electron antineutrinos only . therefore , the number of detected neutrino events can be different from the no - oscillation expectation .
the deficit of observed neutrinos compared to no - oscillation prediction was first detected in a solar neutrino experiment @xcite .
however the rate information alone could not provide sufficient evidence to ascribe conclusively the phenomenon of neutrino `` disappearance '' to flavor oscillations . the energy dependence of neutrino oscillations not only changes the neutrino event rate but also distorts the observed neutrino energy spectrum .
the spectrum distortion provides more information about the pnms matrix components and neutrino mass eigenstates than rate studies alone can .
the inverse beta decay method offers excellent energy sensitivity , which is very valuable for the oscillation studes .
recoil smearing present in this reaction is small compared to detector energy resolution , the latter being the main limiting factor in the accuracy of @xmath8 energy measurement .
other advantages include a relatively large cross section of the reaction and , most importantly , very powerful background suppression due to the characteristic double - coincidence signature . the limitations of this method are the @xmath8 energy threshold of about 1.8 mev and weak directionality .
the success of a neutrino oscillation experiment depends not only on the characteristics of the detector and on the neutrino source but also on the proper choice of the distance between the two ( the baseline ) .
there is no single baseline optimal for all neutrino oscillation studies .
for example , the average baseline of kamland experiment , about 180 km , is fairly good for @xmath9 and especially for @xmath10 but not for @xmath11 , @xmath12 and @xmath13 .
moreover , such parameters as detector resolution , the amount , the nature of the background and the a - priori information about its properties can affect the optimal baseline value . a tunable baseline experiment , which implies movable detector or source , may have a big advantage here .
these considerations , along with the interest in studying terrestrial antineutrinos , led to the idea of a big kamland - like underwater detector @xcite .
the potential of such a detector for neutrino oscillation parameter measurements was the primary motivation for the study presented here .
however , the scope of the actual study is much wider and not limited to the hanohano project .
the results are in fact applicable to any similar medium - baselined experiment .
for baselines associated with current and near - future reactor based neutrino experiments ( up to hundreds of kilometers ) , the matter effects @xcite critical in solar neutrino studies are not significant , so the vacuum oscillation approximation can be used . as was mentioned above
, the inverse beta decay detection is sensitive to electron antineutrinos only .
reactors produce exclusively electron antineutrinos as well , so the observable effect is the apparent `` disappearance '' of a fraction of reactor - produced electron antineutrinos .
the @xmath8 `` survival '' probability is given by the formula @xcite : @xmath14 where @xmath15 . note that `` atmospheric '' mixing angle @xmath16 does not affect the @xmath7 survival and hence not measurable in electron neutrino disappearance experiments .
here , @xmath17 is the `` baseline '' , the distance between the @xmath8 source and the detector .
given the evidence from solar neutrino experiments @xcite that @xmath18 , and the knowledge that @xmath19 from superk @xcite , k2k @xcite , minos @xcite on the one hand and kamland @xcite on the other , only two neutrino hierarchies out of possible six are allowed with currently available data .
they are commonly referred to as `` normal hierarchy , nh '' ( @xmath20 ) and `` inverted hierarchy , ih '' ( @xmath21 ) ( fig . [ hierarchies ] ) .
the former implies that @xmath22 , the latter that @xmath23 , so the sufficiently precise measurement of the those squared mass differences should be enough to establish the neutrino mass hierarchy .
the measurement of @xmath12 , @xmath13 and the mass hierarchy with this approach is possible , in theory , only if @xmath11 is finite and , in practice , if this mixing angle is large enough .
moreover , if the `` solar '' mixing is maximum ( @xmath24 ) , the @xmath12 and @xmath13 become mutually indistinguishable , thus still ruling out the mass hierarchy study , although their values may still be determined without knowing `` which is which '' .
this maximum mixing is strongly disfavored by kamland @xcite and essentially excluded by solar eperiments @xcite .
unfortunately , the same can not be said about the @xmath11 since only the upper limit for this value exists today and there is no experimental evidence that it is not zero .
if it is , then future @xmath8 vacuum oscillation experiments are limited to probing @xmath9 and @xmath10 along with setting still better upper limits on the @xmath11 itself .
that said , global analysis shows a slight preference for non - zero @xmath11 @xcite .
a typical reactor @xmath8 spectrum @xcite multiplied by the inverse beta decay cross section @xcite is shown in figure [ raw ] , dotted .
the antineutrinos are generated in @xmath25 decays of short living fission products of initial fissionable fuel isotopes : @xmath26 , @xmath27 , @xmath28 , @xmath29 . for this study
, the ratio of the isotopes is taken the same as in @xcite .
such a spectrum can be observed at very short - baselined experiments ( baseline @xmath30 1 km ) , where oscillation effects are negligible . at much longer baselines ( 30 km and above ) , the `` solar '' oscillations governed by @xmath9 and @xmath10 lead to an energy - dependent deficit of the observed @xmath8 events .
the effect of those oscillations alone is the `` coarse '' oscillatory pattern of event deficit over the spectrum ( figure [ raw ] , dashed ) with a high amplitude ( determined by @xmath31 ) and a relatively low frequency ( determined by @xmath10 ) .
the amplitude of oscillations driven by the squared mass differences @xmath12 and @xmath13 is proportional to @xmath32 and much smaller than that of `` solar '' oscillations for any currently allowed value of this mixing angle . because @xmath12 and @xmath13 are known to be larger than @xmath10 , the frequency of those sub - dominant oscillations is higher .
a typical @xmath8 energy spectrum expected for a non - zero @xmath11 is shown in figure [ raw ] , solid .
the spectrum analysis approach has already been successfully used by kamland to set by now the best limits on @xmath10 and to confirm sno and superk values for @xmath9 . the idea to measure the remaining three of the five oscillation parameters by precision measurement of the sub - dominant oscillation pattern in a reactor @xmath8 disappearance experiment has been already suggested and thoroughly examined @xcite . in this paper , we examine the capacity of an intermeidate baseline ( 30 - 90 km ) reactor @xmath8 experiment for measuring @xmath9 , @xmath11 , @xmath10 , @xmath12 , @xmath13 and neutrino mass hierarchy .
although this study has been motivated by the project of a big underwater detector hanohano ( @xcite ) , we make no assertions specific for that particular choice .
a special emphasis is placed on the systematic uncertainties and technical limitations present in any real experiment . in our study of the sensitivity to each of the oscillation parameters we take into account the impact of those uncertainties , as well as some detector parameters and the baselines on the resulting performance to formulate in a quantitative way the requirements to which such an experiment must conform .
in this study , we consider the measurement of all the oscillation parameters to which such @xmath8 disappearance experiments are sensitive : @xmath9 , @xmath11 , @xmath10 , @xmath12 , @xmath13 .
three types of detector - related systematic uncertainties are considered which are present to some extent in any experiment and are capable of a non - trivial impact on the sensitivity to the target parameters .
although the success of the borexino experiment @xcite suggests that careful detector design can make the inverse - beta based @xmath8 detection almost background - free , geologically produced antineutrinos @xcite will technically remain a background source for a reactor @xmath8 study in the lower energy region .
what makes this background especially significant is the lack of exact information about its overall intensity and the relative amounts of antineutrinos produced in the `` uranium - radium '' and `` thorium '' decay series .
this amounts to two more systematic uncertainties which have to be left unconstrained within geologically feasible models .
the following detector - related uncertainties were accounted for : * the uncertainty in the predicted event rate .
it is sensitive to the number of target protons ( due to fiducial volume estimation error and uncertainty of the scintillator composition ) , the efficiency of coincidence selection cuts , and live time estimation error . for current similar experiments
, this error tends to be on the order of 1 to 5% .
below we refer to this as `` efficiency '' error .
* the uncertainty in the detector energy resolution estimation .
although often ignored in current experiments , it can have a considerable effect on the measurement of the @xmath11 mixing angle from medium baselines .
numerically , it can be quite big ( about 10% ) depending on the detector calibration options . * the `` linear '' energy scale uncertainty .
this is the uncertainty in the average number of photoelectrons produced by an 1 mev event .
the amount of this uncertainty depends on the detector calibration as well .
normally it can be made quite small ( around 1% ) but its impact on the resulting accuracy of the parameter estimation may still be noticeable .
the energy scale in scintillator - based detectors is in fact substantially non - linear , this non - linearity always producing additional systematic uncertainties which are often rather tricky to parametrize .
however the study of this error is very detector - specific , requires extensive monte - carlo simulations with real calibration data feedback and considering it at this stage would be too speculative . although , kamland internal studies indicate that the nonlinear energy scale uncertainty is less of an issue than the linear one which we can take into account now , hanohano or any other future experiment will have to revisit this issue , once the real experimental feedback from the detector becomes available . the geo - neutrinos yield two more systematic uncertainties : * total detectable terrestrial antineutrino flux , conventionally expressed in terrestrial neutrino units ( tnu ) defined as the number of inverse - beta decay interactions per @xmath33 free protons per year . *
the ratio of @xmath8 originating from the @xmath27 decay chain to those coming from the @xmath34 decay chain .
although geological models do provide some guidelines for the expected geo - neutrino flux and kamland was able to produce the first experimental measurement of the flux , these data are of little use for the purpose of future experiments , including hanohano , because the geo - neutrino estimation precision needed to produce an appreciable advantage over the `` agnostic '' approach is about one order of magnitude higher than available now .
since we ve included background and systematics in the analysis , the direct likelihood approach has been chosen over the combination of the matched digital filter and the fourier transform of the spectrum employed in the earlier publications dedicated to or motivated by hanohano project @xcite .
this approach facilitates the accommodation of the systematic uncertainties and the background .
the likelihood method used here is the unbinned statistical analysis similar to the one employed by kamland experiment @xcite . instead of the real experimental data , a series of `` experiments '' can be simulated as sequences of `` events '' with energies distributed according to the spectra distorted by different oscillation parameters ( including the background ) .
the potential sensitivity is essentially the ability of the data analysis to distinguish between different hypotheses about the oscillation parameter sets .
this study is based on the `` rate+shape '' likelihood function defined for a real or simulated experiment as : @xmath35 where @xmath36 are the event energies , n the number of observed events , @xmath37 the expected number of events ( given the set of parameters @xmath38 ) , @xmath39 @xmath8 energy spectrum normalized to @xmath40 ( after the distortion by the set of parameters @xmath41 ) .
note that while both @xmath42 and @xmath43 are denoted as vectors , these vectors are in different spaces .
the @xmath42 has as many dimensions as the sum of the number of @xmath8 and background events in the experiment , and is fixed for a given experiment .
the @xmath43 lies in the parameter space , its dimensionality being the number of unknown parameters to be fitted , and is variable .
the best fit is obtained by varying the parameter vector @xmath41 to achieve the maximum value of @xmath44 or its logarithm , the latter often being more convenient to calculate and handle .
after the best fit point @xmath45 has been found , the general prescription to evaluate the sensitivity to some individual parameter @xmath46 is the following : * make a small increment ( or decrement ) @xmath47 to the @xmath46 from the `` best fit '' point @xmath48 : @xmath49 . *
find a new point of maximum likelihood by varying all the parameters @xmath41 except for @xmath46 which is kept fixed at @xmath50 .
the new maximum @xmath51 over the subspace constrained by the requirement @xmath52 is not higher than the global maximum @xmath53 .
* repeat the above steps with varying @xmath47 until the condition @xmath54 is met for both positive and negative increments of @xmath47 .
the corresponding points @xmath55 and @xmath56 will limit the confidence range for the k - th parameter .
the value @xmath57 depends on the confidence level for which the range is to be determined . for an individual parameter variation and the confidence level equal to @xmath58 , @xmath59 . in general , for a cl of @xmath60 , @xmath61 .
when instead of a one - dimensional confidence range , a multidimensional confidence region in the parameter subspace is required , the values @xmath57 will be different but the general procedure will not change .
the same is true for the case of discriminating two discrete hypotheses , e.g. between the normal and the inverted neutrino mass hierarchies .
more detailed information on the likelihood analysis can be found in @xcite . for a simulated experiment
, the experimental points @xmath62 do not exist in the first place and are generated according to some reasonable choice of parameters @xmath41 . except for this initial stage ,
the rest of the analysis is the same as described above .
if the initial choice of the parameters to simulate the events is not too far off , this procedure will yield an accurate prediction for the sensitivity of the actual experiment .
systematic uncertainties are introduced by adding `` hidden '' parameters to the parameter space and allowing them to vary during the search for the maximum likelihood as well .
when some information about these values is available a `` penalty '' term is subtracted from @xmath63 to account for the fact that big deviations from the central values of those hidden parameters are unlikely .
if the parameters of uncertainty are normally distributed around their central value and if all systematic uncertainty parameters are uncorrelated , the `` penalty '' term takes on the form : @xmath64 where @xmath65 is the number of systematic uncertainty parameters , @xmath66 is the index of the parameter corresponding to the @xmath67-th uncertainty , @xmath68 is the deviation of the @xmath66-th parameter from its most probable value , and @xmath69 is the value of systematic error ascribed to the @xmath67-th uncertainty .
when some uncertainties are correlated , the penalty term becomes a more general positive definite quadratic form but for our current study that is not the case .
the full equation for the likelihood logarithm with systematic uncertainties takes on the form : @xmath70 in this work , we used the total of nine continuous parameters .
these are four neutrino oscillation parameters : @xmath71 , @xmath72 , @xmath10 , @xmath12 .
note that for each of the two mass hierarchies , @xmath13 is determined by the two other squared mass differences and has not to be introduced into the parameter space .
five parameters were dedicated to systematic uncertainties : two geo - neutrino parameters and three detector - associated systematic errors as described in the previous section .
the geo - neutrino parameters are left unconstrained , which is equivalent to infinite @xmath73 in ( [ penalty - term ] ) .
the default values for the systematic errors in `` efficiency '' , energy resolution estimation and energy scale were taken to be 2% , 8% and 1% , respectively , which is reasonably conservative for experiments of this kind .
additionally , two extreme cases were analyzed : the most `` optimistic '' one with the corresponding parameters fixed at zero deviations as if they were known exactly , and the most `` pessimistic '' one with those three uncertainties left unconstrained as well as geo - neutrino parameters .
although practically impossible , these limiting cases indicate how much sensitivity can be gained by improving the systematics and , conversely , how much would be lost if the systematic errors of the real experiment happen to be worse than expected .
the `` solar '' mixing angle has been fairly well constrained by sno @xcite and kamland @xcite .
our computations suggest that there is still an opportunity for a significant improvement , though .
this measurement is moderately sensitive to detector - based systematic uncertainties but the terrestrial antineutrino background is much more troublesome ( fig [ ss12 ] ) .
these background decrease the sensitivity by about a factor of two and drives the optimum baseline for this parameter to about 60 km , which conflicts with the goal of measuring @xmath11 and @xmath12 in the same experiment as well . constraining the geo - neutrino flux would change the situation , but currently there seems to be no way of doing that . the geo - neutrino flux measurements made by kamland @xcite are not nearly precise enough to improve the situation noticeably and , besides , not directly applicable to a future experiment located elsewhere , especially in the ocean . even with that background and the associated uncertainty , the sensitivity of the medium - baseline experiment is noteworthy . for a 10 kt detector at 50 km from a 6 gwt nuclear plant ,
an exposure of 300 gigawatt - kiloton - years ( 5 years ) is required to achieve the one - sigma confidence range of 0.01 in @xmath74 , which is about 5 times better than the current best estimation . at 60 km
, the same precision can be achieved with just above one - third of this exposure .
although higher energy resolution is always better , the @xmath9 study does not exhibit appreciable dependence on this parameter and @xmath75}$ ] is almost as good as @xmath76}$ ] .
of the detector - associated systematics , the most significant is the `` efficiency '' uncertainty .
the measurement of this parameter by kamland is more difficult to improve on .
for example , the target sensitivity for hanohano is @xmath77 , which would be about three times better than the current best estimation . as our calculations suggest ,
this can be achieved in 300 gigawatt - kiloton - years at the 60 km baseline or in 450 gigawatt - kiloton years from 50 km .
still longer baselines offer better sensitivity for this particular study ( fig [ dms12 ] ) but would be clearly sub - optimal for all other oscillation parameters .
the part ( b ) of the plot shows the dramatic effect of terrestrial neutrino background and , particularly , the uncertainty of this background . without geo - neutinos , the same detector would be four times more efficient at 60 km and seven times at 50 km . the fact that geo - neutrinos drive the optimum baseline towards longer distance may seem somewhat counter - intuitive .
the shorter the baseline , the higher the reactor @xmath8 rate , so the relative fraction of terrestrial @xmath8 background is smaller and should have a smaller effect .
however , at shorter baselines , the reactor @xmath8 deficit due to oscillations appears mostly in the lower - energy zone where it is harder to separate from the variation in the terrestrial neutrino background . like the @xmath9 measurement
, this study is not demanding of detector energy resolution and not particularly sensitive to detector - associated systematics .
this is a very important oscillation parameter not only because of its theoretical significance but also because its value defines the amplitude of the sub - dominant high - frequency oscillations governed by @xmath78 and @xmath79 .
only if @xmath11 is not zero ( and not too small ) is it possible to measure those mass squared differences in @xmath8 disappearance experiments .
currently , only an upper bound for this angle is known ( from the chooz experiment @xcite ) : @xmath80 .
several experiments are proposed or already under construction to set better limits and the hanohano detector can contribute to those efforts .
the sensitivity profiles ( figure [ ss13 ] ) show that medium baselines ( above 30 km ) are not optimal for this study and much shorter ones are better from the statistical standpoint
. however even at 50 km the absolute sensitivity can be quite impressive with a big detector . except for the longest baselines ( 60 km and above ) which are clearly sub - optimal
, this study is not severly affected by the geo - neutrino background and its uncertainties .
the systematics of the detector itself , however , play a more important role here . at 50 km ,
the main systematic error is the uncertainty of energy resolution estimation , followed by the `` efficiency '' error . at shorter baselines the `` efficiency '' uncertainty dominates .
although the medium baselines have a strong statistical disadvantage for @xmath11 measurement , they also have the compelling feature that systematic uncertainties do not ruin the measurement . unlike the shorter baseline experiments where
relatively more information is obtained through the neutrino event rate , the spectrum shape distortion characteristic of medium baseline is not so easy to imitate by any of the detector systematic errors .
this means that , in the long run when even better accuracy for @xmath11 is required , medium baseline experiments may prove to be more robust .
figure [ ss13r ] exhibits another important feature of this measurement : its energy resolution dependence . although not as critical as for the hierarchy study ( below ) , the effect of detector energy resolution is quite noticeable . compromising this parameter to @xmath75}$ ] from the @xmath76}$ ] ( as projected for hanohano )
will cost about 2.5 times the exposure .
unlike all previously considered parameters where the potential sensitivity of an experiment could be predicted more or less accurately based just on the @xmath8 exposure , this measurement depends on the value of @xmath11 which is still unknown .
any quantitative sensitivity prediction makes sense only with some particular value of @xmath11 in mind .
the larger the mixing angle , the easier it is to determine @xmath12 , @xmath13 and neutrino mass hierarchy .
it has been found that the sensitivity scales approximately as the square of @xmath32 . in other words , getting the same sensitivity in @xmath12 , @xmath13 and neutrino mass hierarchy if @xmath81 will take four times the exposure required if @xmath82 . in this paper we carried out calculations for two scenarios : @xmath83 , and @xmath84 .
another ambiguity associated with the study of @xmath12 and @xmath13 follows from the closely related question of neutrino mass hierarchy .
depending on the actual value of the mixing angle @xmath11 , the neutrino mass hierarchy may turn out unfeasible to establish at an adequate cl . at the same time , within each of the two possible hierarchies , stringent limits on both @xmath12 and @xmath13 may still be set with reasonable exposures .
since the ambiguity is at worst only a two - fold one , it makes sense to estimate the `` known - hierarchy '' sensitivity to either @xmath12 or @xmath13 .
figure [ dms13 - 50 ] and [ dms13 - 25 ] show that shorter baselines are better for this study , although this trend is not as pronounced as with @xmath11 measurement and actually reverses below 25 km .
it is clear that @xmath12 study is not systematics - constrained , including the systematics from geo - neutrinos . on the other hand ,
the dependence on energy resolution for this measurement is even stronger than that for @xmath11 ( figure [ dms13r-50 ] and [ dms13r-25 ] ) .
like the @xmath12/@xmath13 measurement , the hierarchy study depends on the actual value of @xmath11 .
our calculations suggest that for any @xmath11 it takes more statistics to make a high confidence level conclusion about the hierarchy than to measure @xmath12 and @xmath13 to an accuracy of @xmath85 .
this makes reliable hierarchy determination with a 10 kt detector feasible only if the mixing angle turns out to be quite high ( @xmath86 ) .
the sensitivity dependence on this value is approximately quadratic in the exposure as well .
the baseline profile of the sensitivity to the hierarchy is shown in figure [ hier ] .
after taking into account the geo - neutrino background and the uncertainties the optimum baseline remains in the same range as was found earlier with less comprehensive models @xcite 50 km or slightly more . of the systematic errors ,
the most damaging is the geo - neutrino flux uncertainty , although its effect is not as decisive as for the solar parameter studies .
the hierarchy study proves to be the most demanding of the detector energy resolution ( figure [ hier - r ] ) . even within the best values for that parameter of the detector achievable today the sensitivity dependence on
the energy resolution enters its asymptotic 4-th power curve .
in particular , this implies that between two detectors of the same photocathode area but different volumes , the bigger detector will offer _ inferior _ sensitivity : all other parameters being equal , a smaller relative photocathode coverage will lead to lower resolution which will prevail over the higher reactor @xmath8 statistics .
theoretically , the mass hierarchy study is secondary to the measurement of @xmath12 and @xmath13 and is determined immediately after those mass differences are found .
the analysis of the oscillated @xmath8 energy spectrum yields all mass differences ( provided , @xmath87 and @xmath88 ) . in practice , however , the squared mass differences can be measured with limited accuracy only and at a limited cl
. this may not be sufficient to determine the hierarchy .
moreover , it has been found that for any combination of @xmath89 and @xmath90 there exists another one ( denoted below as @xmath91 and @xmath92 ) that delivers a similar oscillation pattern despite comprising the oppositie mass hierarchy .
the similarity is never perfect and , given enough statistics , it is always possible to distinguish between the two spectra but it may take much more exposure to discriminate between those `` conjugate '' opposite hierarchy solutions than to constrain the squared mass differences within one of the solutions with a remarkable precision . in figure [ likedef ] ,
the curves provide the measure of relative `` unlikeliness '' of an alternative hypothesis , assuming normal hierarchy and @xmath93 , for which the experiment was simulated .
zero @xmath94 means an indistinguishable hypothesis , the higher its value , the less statistics is needed to discriminate the hypothesis .
solid lines show the normal hierarchy and dotted lines show the inverted one .
the dashed vertical lines through the centers of the dotted curves point to the @xmath95 which combined with the inverted hierarchy provide the closest similarity to the simulated physical spectrum .
comparison of figure [ likedef](a ) and figure [ likedef](b ) explains why the 60 km baseline offers better hierarchy discrimination than 40 km , although the latter yields significantly better sensitivity to @xmath12/@xmath13 within each of the two hierarchies : the vertex of the quasi - parabolic dotted curve is located higher for 60 km .
the `` conjugate '' @xmath95 for a given real @xmath95 is not the same for different baselines , which implies that two measurements at different baselines may offer improved efficiency . for instance , the allowable values of @xmath95 to which the 60 km baselined measurement is least sensitive are much better excluded at 40 km and vice versa . in case of a land - based detection , a multiple - detector configuration can be considered .
hanohano , additionally , can use the advantage of its movability and make two consecutive exposures instead of one twice as long . indeed , as the comparison between figure [ likedef](c ) and figure [ likedef](d ) shows , the combination of 60 and 40 km baselined observations should provide a better hierarchy resolution than one twice as long at the practically optimal 50 km .
although the advantage is marginal , we considered only a two - baseline combination with equal exposures .
a more systematic optimization with different exposures and possibly more than just two baselines should offer further gains .
with a single detector and a 300 kiloton - gigawatt - year exposure and a 5 - 6 gw thermal power @xmath8 source , both @xmath10 and @xmath9 are expected to be measured with the accuracy of 1% , a three to five time improvement on the current best limits @xcite .
an experiment with a more powerful reactor @xmath8 source but the same exposure due to a smaller detector size or a shorter livetime will have a slight advantage because of the relatively smaller geo - neutrino effect .
detector - associated systematic uncertainties do not appear to be a significant limiting factor in the resulting accuracy .
these studies are not particularly sensitive to the detector resolution and @xmath96}$ ] would be quite sufficient .
however the expected sensitivity , especially that for @xmath10 , will be severely handicapped by the presence of geo - neutrinos and in particular by the lack of accurate estimation for the intensity of this background source , effectively turning its value into yet another systematic uncertainty the dominating one for this study .
this is not a `` hard '' limitation since geo - neutrinos , having a different spectrum , can not mimic the reactor @xmath8 oscillaton pattern , but rather an efficiency impairment .
the study of the @xmath11 mixing angle is different , in that the medium baselines that this study deals with are not optimal . on a per - event basis ,
the baseline dependence of sensitivity is rather flat , but short baselines gives quadratically higher event rate and hence lower statistical error . on the other hand , longer baselines offer more robustness with respect to the `` efficiency '' systematic uncertainty . given the careful design of hanohano or any similar ( considering the size and the baseline ) detector and its accurate calibration , it is still possible to reach an accuracy of better than 0.02 in @xmath97 evaluation , which is competitive with the dedicated experiments like double chooz and daya bay .
the sensitivity to this mixing angle does not exhibit strong dependence on its own real value .
in other words , setting the upper limit of 0.02 for @xmath32 if the angle happens to be zero takes about as much exposure as setting the range between 0.03 and 0.07 if the value is really 0.05 .
the measurement of @xmath12 and @xmath13 and the closely related question of neutrino mass hierarchy are common in their dependence on the actual value of @xmath11 . if the angle turns out to be big enough ( within currently allowed values )
, then some spectacular results are possible .
if it is zero or very small , nothing interesting can be measured with either haonhano or any other similar experiment .
the necessary exposure is approximately inversely quadratic to the value of @xmath32 in both cases , although the hierarchy measurement generally requires much more statistics . for a moderately optimistic scenario in which @xmath98 , hanohano or a similar experiment can yield a very good estimation for the values of the squared mass differences but
reliable mass hierarchy separation may call for prohibitively long exposures .
another feature that these measurements have in common is the requirement for the excellent energy resolution . generally , assuming either normal or inverted hierarchy , the problem of @xmath12 and @xmath13 becomes simpler , less demanding of the energy resolution and with higher chance of success for unfavorably small values of @xmath11
. even if the hierarchy question is not conclusively answered at that stage , the squared mass differences can still be measured , even though the remaining hierarchy ambiguity will split the allowable solutions into two groups .
in such a case , the result of the @xmath12 and @xmath13 study will have the form : @xmath99 @xmath100 for normal hierarchy , and @xmath101 @xmath102 for inverted hierarchy , with one solution somewhat more favored over the other ( i.e. should the hierarchy discrimination be achieved to marginal confidence levels ) . here
@xmath103 and @xmath104 are best fit values for the @xmath12 for the normal and inverted scenarios , respectively , the @xmath10 is expected to be found from the same experiment with superior accuracy and @xmath105 and @xmath106 are the error bars for both the @xmath12 and @xmath13 in the normal and inverted hierarchy , respectively .
as was proposed in @xcite and discussed in numerous later publications , neutrinos may have non - standard interactions which could affect the flavor content at the source and also the flavor content detected . for our case at hand , it means that the observed mixing angles @xmath9 and @xmath11 in general will differ from the `` true '' mixing angles .
for example , the measured @xmath11 can be larger than the true @xmath11 @xcite .
since the survival probability ( [ 3flavor ] ) depends on the effective @xmath11 , the nsi have no adverse effect on the determination of @xmath12 , @xmath107 and the neutrino mass hierarchy .
in fact an effective @xmath11 larger than the real one will be advantageous for these studies . at the same time
, the effective @xmath11 measured at different baselines are going to be different , should these interactions take place . this way , medium baseline experiments targeting @xmath11 will become complementary to the short baseline ones in testing new physics .
* medium - baselined @xmath8 oscillation experiments are not systematics - constrained .
this follows from the shapes of oscillated spectra .
in particular , physically feasible systematic errors do not tend to imitate the spectral distortions characteristic of the neutrino oscillations .
* there is no single baseline optimal for all oscillation studies .
the difference in sensitivity profiles is big enough to give an advantage to multiple detector or / and movable detector configurations . even in individual studies
where a pronounced baseline optimum exists , a multiple baseline configuration can outperform a single baseline configuration , as has been shown for the neutrino mass hierarchy discrimination case .
s. t. petcov and m. piai , phys .
lett b. * 533 * , 94 ( 2002 ) arxiv : hep - ph/0112074 .
s. choubey , s. t. petcov , m. piai , phys .
d * 68 * , 113006 ( 2003 ) arxiv : hep - ph/0306017v1 .
a. bandyopadhyay , s. choubey , s. goswami , s.t .
petcov , d.p .
roy , ( 2008 ) arxiv:0804.4857v1 [ hep - ph ] | despite the dramatic progress made in neutrino oscillation studies recently , several fundamental neutrino parameters remain either unknown or poorly measured .
we discuss in detail a method for their measurement by precision studies of oscillation - caused neutrino energy spectrum distortions , for which a large underwater inverse beta decay detector appears to be a perfect tool .
results determine optimal baselines and necessary exposures in the presence of systematic uncertainties and the unavoidable background from terrestrial antineutrinos . |
the work of wyner @xcite led to the development of the notion of secrecy capacity , which quantifies the maximum rate at which a transmitter can reliably send a secret message to a receiver , without an eavesdropper being able to decode it .
more recently , researchers have considered secrecy for the two - user broadcast channel , where each receiver acts as an eavesdropper for the independent message transmitted to the other .
this problem was addressed in @xcite , where inner and outer bounds for the secrecy capacity region were established .
further work in @xcite studied the multiple - input single - output ( miso ) gaussian case , and @xcite considered the general mimo gaussian case .
it was shown in @xcite that , under an input covariance constraint , both confidential messages can be simultaneously communicated at their respective maximum secrecy rates , where the achievablity is obtained using secret dirty - paper coding ( s - dpc ) . however , under an average power constraint , a computable secrecy capacity expression for the general mimo case has not yet been derived . in principle , the secrecy capacity for this case could be found by an exhaustive search over the set of all input covariance matrices that satisfy the average power constraint @xcite . clearly , the complexity associated with such a search and the implementation of dirty - paper encoding and decoding make such an approach prohibitive except for very simple scenarios , and motivates the study of simpler techniques based on linear precoding .
while low - complexity linear transmission techniques have been extensively investigated for the broadcast channel ( bc ) without secrecy constraints , e.g. , @xcite-@xcite , there has been relatively little work on considering secrecy in the design of linear precoders for the bc case . in @xcite , we considered linear precoders for the mimo gaussian broadcast channel with confidential messages based on the generalized singular value decomposition ( gsvd ) @xcite .
it was shown numerically in @xcite that , with an optimal allocation of power for the gsvd - based precoder , the achievable secrecy rate is very close to the secrecy capacity region . in this paper , we show that for a two - user mimo gaussian bc with arbitrary numbers of antennas at each node and under an input covariance constraint , linear precoding is optimal and achieves the same secrecy rate region as s - dpc for certain input covariance constraints , and we derive an expression for the optimal precoders in these scenarios .
we then use this result to develop a sub - optimal closed - form algorithm for calculating linear precoders for the case of average power constraints .
our numerical results indicate that the secrecy rate region achieved by this algorithm is close to that obtained by the optimal s - dpc approach with a search over all suitable input covariance matrices . in section [ secii ]
, we describe the model for the mimo gaussian broadcast channel with confidential messages and the optimal s - dpc scheme , proposed in @xcite . in section [ seciii ]
, we consider a general mimo broadcast channel under a matrix covariance constraint , we derive the conditions under which linear precoding is optimal and achieves the same secrecy rate region as s - dpc , and we find the corresponding optimal precoders .
we then present our sub - optimal algorithm for designing linear precoders for the case of an average power constraint in section [ seciv ] , followed by numerical examples in section [ secv ] .
section [ secvi ] concludes the paper .
* notation : * vector - valued random variables are written with non - boldface uppercase letters ( _ e.g. , _ @xmath0 ) , while the corresponding non - boldface lowercase letter ( @xmath1 ) denotes a specific realization of the random variable .
scalar variables are written with non - boldface ( lowercase or uppercase ) letters .
the hermitian ( i.e. , conjugate ) transpose is denoted by @xmath2 , the matrix trace by tr ( . ) , and * i * indicates an identity matrix .
the inequality @xmath3 ( @xmath4 ) means that @xmath5 is hermitian positive ( semi-)definite .
mutual information between the random variables @xmath6 and @xmath7 is denoted by @xmath8 , @xmath9 is the expectation operator , and @xmath10 represents the complex circularly symmetric gaussian distribution with zero mean and variance @xmath11 .
we consider a two - receiver multiple - antenna gaussian broadcast channel with confidential messages , where the transmitter , receiver 1 and receiver 2 possess @xmath12 , @xmath13 , and @xmath14 antennas , respectively .
the transmitter has two independent confidential messages , @xmath15 and @xmath16 , where @xmath15 is intended for receiver 1 but needs to be kept secret from receiver 2 , and @xmath16 is intended for receiver 2 but needs to be kept secret from receiver 1 @xcite .
the signals at each receiver can be written as : @xmath17 where @xmath1 is the @xmath18 transmitted signal , and @xmath19 is white gaussian noise at receiver @xmath20 with independent and identically distributed entries drawn from @xmath21 .
the channel matrices @xmath22 and @xmath23 are assumed to be unrelated to each other , and known at all three nodes .
the transmitted signal is subject to an average power constraint when @xmath24 for some scalar @xmath25 , or it is subject to a matrix power constraint when @xcite : @xmath26 where @xmath27 is the transmit covariance matrix , and @xmath28 .
compared with the average power constraint , ( [ lin3 ] ) is rather precise and inflexible , although for example it does allow for the incorporation of per - antenna power constraints as a special case .
it was shown in @xcite that for any jointly distributed @xmath29 such that @xmath30 forms a markov chain and the power constraint over @xmath0 is satisfied , the secrecy rate pair @xmath31 given by @xmath32 is achievable for the mimo gaussian broadcast channel given by ( [ lin1 ] ) , where the auxiliary variables @xmath33 and @xmath34 represent the precoding signals for the confidential messages @xmath15 and @xmath16 , respectively @xcite . in @xcite , the achievablity of the rate pair ( [ lin4 ] ) was proved .
liu _ et al . _
@xcite analyzed the above secret communication problem under the matrix power - covariance constraint ( [ lin3 ] ) .
they showed that the secrecy capacity region @xmath35 is rectangular .
this interesting result implies that under the matrix power constraint , both confidential messages @xmath15 and @xmath16 can be _ simultaneously _ transmitted at their respective maximal secrecy rates , as if over two separate mimo gaussian wiretap channels . to prove this result , liu _
_ showed that the secrecy capacity of the mimo gaussian wiretap channel can also be achieved via a coding scheme that uses artificial noise and random binning ( * ? ? ? * theorem 2 ) . under the matrix power constraint ( [ lin3 ] ) , the achievablity of the optimal corner point @xmath36 given by ( * ? ? ? *
theorem 1 ) @xmath37 is obtained using dirty - paper coding based on double binning , or as referred to in @xcite , secret dirty paper coding ( s - dpc ) .
more precisely , let @xmath38 maximize ( [ lin5 ] ) , and let @xmath39 where @xmath40 and @xmath41 are two independent gaussian vectors with zero means and covariance matrices @xmath42 and @xmath43 , respectively , and the precoding matrix @xmath44 is defined as @xmath45 .
one can easily confirm the achievablity of the corner point @xmath36 by evaluating ( [ lin4 ] ) for the above random variables and noting that in ( [ lin1 ] ) , @xmath46 . note that under the matrix power constraint @xmath47 , the input covariance matrix that achieves the corner point in the secrecy capacity region satisfies @xmath48 @xcite .
the matrix @xmath49 that maximizes ( [ lin5 ] ) is given by @xcite @xmath50\bc^h \bs^{\frac{1}{2}}\end{aligned}\ ] ] where @xmath51 $ ] is an invertible is invertible since both components of the pencil ( [ lin9 ] ) are positive definite . ]
generalized eigenvector matrix of the pencil @xmath52 satisfying @xcite @xmath53\bc=\mathbf{\lambda}\\ & \bc^h\left[\bs^{\frac{1}{2}}\bg^h\bg\bs^{\frac{1}{2}}+\bi\right]\bc=\bi \ ; , \end{split}\end{aligned}\ ] ] where @xmath54 contains the generalized eigenvalues sorted without loss of generality such that @xmath55 the quantity @xmath56 denotes the number of generalized eigenvalues greater than one @xmath57 , and defines the following matrix partitions : @xmath58 \qquad \bc = [ \bc_1\ ; \,\bc_2 ] \ ; , \label{lin11}\ ] ] where @xmath59 , @xmath60 , @xmath61 contains the @xmath56 generalized eigenvectors corresponding to @xmath62 and @xmath63 the @xmath64 generalized eigenvectors corresponding to @xmath65 .
now , by applying ( [ lin8 ] ) in ( [ lin5 ] ) , the corner rate pair @xmath36 can be calculated as ( ( * ? ? ? * theorem 3 ) ) @xmath66 for the average power constraint in ( [ lin2 ] ) , there is no computable secrecy capacity expression for the general mimo case . in principle the secrecy capacity region for the average power constraint , @xmath67 ,
could be found through an exhaustive search over all suitable matrix power constraints @xcite , ( * ? ? ?
* lemma 1 ) : @xmath68 for any given semidefinite @xmath47 , @xmath35 can be computed as given by ( [ lin13 ] ) . then , the secrecy capacity region @xmath67 is the convex hull of all of the obtained corner points using ( [ lin13 ] ) .
the complexity associated with such a search , as well as that required to implement dirty - paper encoding and decoding , are the main drawbacks of using s - dpc to find the secrecy capacity region @xmath67 for the average power constraint .
this makes linear precoding ( beamforming ) techniques an attractive alternative because of their simplicity . to address the performance achievable with linear precoding
, we first describe the conditions under which linear precoding is optimal in attaining the same secrecy rate region that is achievable via s - dpc , when the broadcast channel is under an input covariance constraint .
in particular , in the next section we show that this equivalence holds for matrix power constraints that satisfy a certain property , and we derive the linear precoders that achieve optimal performance .
section [ seciv ] then uses these results to derive a sub - optimal algorithm for the case of the average power constraint .
in this section we answer the following questions : a. for a given general mimo gaussian bc described by ( [ lin1 ] ) , where each node has an arbitrary number of antennas and the channel input is under the covariance constraint ( [ lin3 ] ) , is there any @xmath69 for which linear precoding can attain the secrecy capacity region ?
b. if yes , how can such @xmath47 be described ?
c. for such @xmath47 , what is the optimal linear precoder that allows the rectangular s - dpc capacity region given by ( [ lin13 ] ) to be achieved ? d. if @xmath47 does not satisfy the condition for optimal linear precoding in ( a ) , what is the worst - case loss in secrecy capacity incurred by using the linear precoding approach described in (
b ) anyway ? to begin , we give the following theorem as an answer to questions ( a ) and ( b ) above .
[ lin_thm1 ] suppose the matrix power constraint @xmath69 on the input covariance @xmath27 in ( [ lin3 ] ) leads to generalized eigenvectors in ( [ lin10 ] ) that satisfy @xmath70 , i.e. @xmath71 . then the secrecy capacity region @xmath35 can be achieved with @xmath72 , where @xmath33 and @xmath34 are _ independent _ gaussian precoders respectively corresponding to @xmath15 and @xmath16 , with zero means and covariance matrices @xmath73 and @xmath74 , with @xmath73 defined in ( [ lin8 ] ) .
recall that for any @xmath69 , the secrecy capacity region @xmath35 is rectangular , so we only need to show that when @xmath75 , the linear precoders @xmath33 and @xmath34 characterized in this theorem are capable of achieving the corner point @xmath76 given by ( [ lin13 ] ) . from ( [ lin4 ] ) , the achievable secrecy rate @xmath77 is given by @xmath78 where ( [ lin16 ] ) and the second part of ( [ lin17 ] ) come from the fact that @xmath33 and @xmath34 are independent
. equation ( [ lin18 ] ) is proved in appendix a. one can similarly show that @xmath79 is achievable to complete the proof .
theorem [ lin_thm1 ] shows that the secrecy capacity region corresponding to any @xmath47 with orthogonal @xmath61 and @xmath63 can be achieved using either linear independent precoders @xmath33 and @xmath34 , as defined in theorem [ lin_thm1 ] , or using the s - dpc approach , as given by ( [ lin6 ] ) .
the next theorem expands on the answer to question ( b ) above , and also addresses ( c ) .
first however we present the following lemma which holds for any @xmath69 .
[ lin_lem1 ] for a given bc under the matrix power constraint ( [ lin3 ] ) , for any @xmath69 we have @xmath80 , where @xmath81 is the number of positive eigenvalues of the matrix @xmath82
. please see appendix b. the following theorem presents a more specific condition on @xmath47 that results in generalized eigenvectors that satisfy @xmath70 .
[ lin_thm2 ] for any @xmath69 , the generalized eigenvectors @xmath61 and @xmath63 in ( [ lin10 ] ) are orthogonal _ iff _ there exists a matrix @xmath83 such that @xmath84 and @xmath85 simultaneously block diagonalizes @xmath86 and @xmath87 : @xmath88 \qquad \bt^h\bg^h\bg\bt = \left[\begin{array}{ccc}\bk_{\bg1 } & \b0\\ \b0 & \bk_{\bg2}\end{array}\right ] \;,\ ] ] where the @xmath89 matrices @xmath90 and @xmath91 satisfy @xmath92 and @xmath93 .
the proof begins by noting that if @xmath84 , then the pencil in ( [ lin9 ] ) and @xmath94 have exactly the same generalized eigenvalue matrix @xmath95 , and thus the same secrecy capacity regions .
the remainder of the proof can be found in appendix c. while algorithms exist to find @xmath85 that jointly block diagonalizes @xmath86 and @xmath87 ( see for example @xcite and references therein ) , as mentioned in appendix c only those @xmath85 that lead to @xmath96 and @xmath97 are acceptable .
later , we will demonstrate that for any bc there are an infinite number of matrix constraints @xmath47 that can achieve such a block diagonalization and hence allow for an optimal linear precoding solution . to conclude this section ,
we now answer question ( d ) posed above .
define the projection matrices @xmath98 and @xmath99 , and note that in general , equation ( [ lin8 ] ) is equivalent to @xmath100 and @xmath101 .
when @xmath102 , the optimal covariance matrices for @xmath33 and @xmath34 also satisfy @xmath103 the following theorem explains the loss in secrecy that results when linear precoding with these covariances is used for a matrix constraint @xmath47 that does _ not _ satisfy @xmath102 .
[ lin_lem2 ] assume a linear precoding scheme @xmath72 for independent gaussian precoders @xmath33 and @xmath34 with zero means and covariance matrices @xmath104 and @xmath105 , respectively .
also define @xmath106 .
the loss in secrecy capacity that results from using this approach in the two - user bc is _ at most _
@xmath107 for each user .
in particular , the following secrecy rate pair is achievable : @xmath108 see appendix d. [ lin_rem2 ] note that if @xmath61 and @xmath63 are orthogonal , then @xmath109 and @xmath110 , @xmath111 , is achievable , as discussed in theorem 1 .
so far we have shown that if the broadcast channel ( [ lin1 ] ) is under the matrix power constraint @xmath47 ( [ lin3 ] ) , then linear precoding as defined by theorem [ lin_thm1 ] is an optimal solution when @xmath47 satisfies the condition described in theorem [ lin_thm2 ] . in the following
we propose a suboptimal closed - form linear precoding scheme for the general mimo gaussian bc under the _ average _ power constraint ( [ lin2 ] ) , where as mentioned earlier there exists no optimal closed - form solution that characterizes the secrecy capacity region .
we begin with some preliminary results , then we develop the algorithm for the general mimo case , and finally we present an alternative algorithm specifically for the miso case since it offers additional insight .
[ lin_remetx1 ] suppose that the input covariance matrix @xmath27 leads to a point on the pareto boundary of the secrecy capacity region given by ( [ lin14 ] ) under the average power constraint ( [ lin2 ] ) .
then @xmath112 and @xmath27 can not have any component in the nullspace of @xmath113 , and thus @xmath114 , where @xmath115 , @xmath116 and @xmath117 contains the singular vectors corresponding to the non - zero singular values of @xmath113 . according to remark [ lin_remetx1 ]
, we can assume without loss of generality that @xmath113 is full - rank ; otherwise , we could replace @xmath118 with @xmath119 and have an equivalent problem where @xmath120 is full - rank and the secrecy capacity region is the same ( in such a case , @xmath12 would then represent the number of transmitted data streams rather than the number of antennas ) . with this result
, we have the following lemma . [ lin_lem3 ] define @xmath121 then @xmath122 and @xmath123 commute and hence share the same set of eigenvectors : @xmath124 where @xmath125 is the ( unitary ) matrix of eigenvectors and @xmath126 the corresponding eigenvalues .
see appendix e. without loss of generality , we assume that the columns of @xmath125 are sorted such that the first @xmath127 diagonal elements of @xmath128 are greater than the first @xmath127 diagonal elements of @xmath129 , and the last @xmath130 diagonal elements of @xmath128 are less than or equal to those of @xmath129 . recall from lemma [ lin_lem1 ] that @xmath131 , where @xmath81 is the number of positive eigenvalues of @xmath82 .
thus , @xmath132 \qquad \bsig_2 = \left[\begin{array}{ccc}\bsig_{2\rho } & \b0\\\b0 & \bsig_{2\bar{\rho}}\end{array}\right]\ ] ] where @xmath133 is @xmath134 , @xmath135 is @xmath136 , @xmath137 and @xmath138 .
now define @xmath139 where @xmath140 and @xmath125 are given in lemma [ lin_lem3 ] and @xmath141 is any block - diagonal matrix partitioned in the same way as @xmath128 and @xmath129 . with these definitions ,
we see from ( [ lin21 ] ) that @xmath142 and @xmath143 are block diagonal .
thus , from theorem [ lin_thm2 ] , a bc with the matrix power constraint @xmath144 leads to a matrix pencil @xmath145 with generalized eigenvectors @xmath146 $ ] that satisfy @xmath147 , where @xmath148 correspond to generalized eigenvalues that are larger or less - than - or - equal - to one , respectively .
[ lin_remetx2 ] since the above result holds for any block - diagonal @xmath149 with appropriate dimensions , then for every bc there are an infinite number of matrix power constraints @xmath150 that achieve a block diagonalization and hence allow for an _ optimal _ linear precoding solution . in the following ,
we restrict our attention to diagonal rather than block - diagonal matrices @xmath151 , for which a closed form solution can be derived . from theorem [ lin_thm1 ] , we have the following result . [ lin_lem4 ] for any diagonal @xmath152 , the secrecy capacity of the broadcast channel in ( [ lin1 ] ) under the matrix power constraint @xmath153 defined in ( [ lin20])-([lin23 ] ) can be obtained by linear precoding . in particular , @xmath154 = v_1 + v_2\end{aligned}\ ] ] where @xmath155 and @xmath156 are independent gaussian random vectors with zero means and covariance matrices @xmath157 and @xmath158 such that @xmath159 \ ; , \end{aligned}\ ] ] and as before @xmath160 represent independently encoded gaussian codebook symbols corresponding to the confidential messages @xmath15 and @xmath16 , with zero means and covariances @xmath161 and @xmath162 respectively given by @xmath163 \bphi_{\bw}^h \bw \label{lin25a } \\
\bs_\bw - \bk_{t\bw}^ * & = \bw\bphi_{\bw}\left [ \begin{array}{cc } \b0 & \b0 \\ \b0 & \bp_2 \end{array } \right ] \bphi_{\bw}^h \bw \ ; . \label{lin25b}\end{aligned}\ ] ] the matrix @xmath164 simultaneously block diagonalizes @xmath86 and @xmath87 , so by theorems [ lin_thm1 ] and [ lin_thm2 ] we know that linear precoding can achieve the secrecy capacity region .
the proof is completed in appendix f by showing the equality in ( [ lin25 ] ) , and showing that ( [ lin25a ] ) corresponds to the optimal covariance in ( [ lin8 ] ) . from the proof in appendix f and ( [ lin21])-([lin23 ] ) , we see that under the matrix power constraint @xmath150 given by ( [ lin24 ] ) with diagonal @xmath151 , the general bc is transformed to an equivalent bc with a set of parallel independent subchannels between the transmitter and the receivers , and it suffices for the transmitter to use independent gaussian codebooks across these subchannels . in particular , the diagonal entries of @xmath157 and @xmath158 represent the power assigned to these independent subchannels prior to application of the precoder @xmath165 in ( [ lin23 ] ) do not represent the actual transmitted power , since the columns of @xmath165 are not unit - norm . ] . from ( [ lin25 ] ) ,
the signals at the two receivers are given by @xmath166 + \bz_1 \\ & = & \bgam_1 \bsig_1 \left [ \begin{array}{c } \bv'_1 \\ \bv'_2 \end{array } \right ] + \bz_1 \\ & = & \bgam_1 \left [ \begin{array}{c } \bsig_{1\rho } \bv'_1 \\ \bsig_{1\bar{\rho } } \bv'_2 \end{array } \right ] + \bz_1 \\ \by_2 & = & \bgam_2 \left [ \begin{array}{c } \bsig_{2\rho } \bv'_1 \\ \bsig_{2\bar{\rho } } \bv'_2 \end{array } \right ] + \bz_2 \ ; , \end{aligned}\ ] ] where @xmath167 are unitary .
the confidential message for receiver 1 is thus transmitted with power loading @xmath157 over those subchannels which are degraded for receiver 2 ( @xmath168 ) , while receiver 2 s confidential message has power loading @xmath158 over subchannels which are degraded for receiver 1 ( @xmath169 ) .
any subchannels for which the diagonal elements of @xmath170 are equal to those of @xmath171 are useless from the viewpoint of secret communication , but could be used to send common non - confidential messages . from theorem [ lin_thm1 ] , the rectangular secrecy capacity region of the mimo gaussian bc ( [ lin1 ] ) under the matrix power constraint @xmath150 ( [ lin24 ] ) is defined by the corner points @xmath172 where @xmath173 is given by ( [ linap35 ] ) in appendix f. note that we have explicitly written @xmath174 as a function of the diagonal matrix @xmath175 to emphasize that @xmath157 contains the only parameters that can be optimized for @xmath174 .
more precisely , since for a given matrix power constraint @xmath150 , @xmath176 and @xmath177 are channel dependent and thus fixed , as shown in ( [ lin21])-([lin22 ] ) .
a similar description is also true for @xmath178 .
here we propose our sub - optimal closed form solution based on linear precoding for the broadcast channel under the _ average _ power constraint ( [ lin2 ] ) .
the goal is to find the diagonal matrix @xmath151 in ( [ lin24 ] ) that maximizes @xmath179 in ( [ lin27 ] ) for a given allocation of the transmit power to message @xmath180 , and that satisfies the average power constraint in ( [ lin28 ] ) . ]
@xmath181 noting that @xmath125 can be written as @xmath182 $ ] , where @xmath183 is a @xmath184 submatrix corresponding to the eigenvalues in @xmath176 , ( [ lin28 ] ) can be rewritten as @xmath185 where we defined positive definite matrices @xmath186 , @xmath111 .
our sub - optimal closed - form solution for the bc under the average power constraint ( [ lin2 ] ) is not optimal , since instead of doing an exhaustive search over all @xmath69 with @xmath187 as indicated in ( [ lin14 ] ) , we will only consider specific @xmath47 matrices of the form given for @xmath150 in ( [ lin24 ] ) with diagonal @xmath151 .
since @xmath188 is only a function of @xmath189 , @xmath190 and @xmath191 can be optimized separately for any power fraction @xmath192 ( @xmath193 ) under the constraints @xmath194 and @xmath195 , respectively .
[ lin_lem5 ] for any @xmath192 , @xmath193 , the diagonal elements of the optimal @xmath196 and @xmath197 are given by @xmath198 where @xmath199 , @xmath200 , and @xmath201 are the @xmath202 diagonal elements of @xmath176 , @xmath177 , and @xmath203 , respectively , where @xmath204 .
also @xmath205 , @xmath206 , and @xmath207 are the @xmath202 diagonal elements of @xmath171 , @xmath170 , and @xmath208 , respectively , where @xmath209 . the lagrange parameters @xmath210 and @xmath211 are chosen to satisfy the average power constraints @xmath194 and @xmath195 , respectively .
we want to optimize diagonal matrices @xmath157 and @xmath158 so that the secrecy rates @xmath190 and @xmath191 , given by ( [ lin27 ] ) , are maximized for a given @xmath192 , @xmath193 .
since @xmath212 only depends on @xmath189 , the two terms in ( [ lin27 ] ) can be maximized independently .
we show the result for @xmath213 ; the procedure for @xmath214 is identical . from ( [ lin27 ] ) , the lagrangian associated with @xmath215 is @xmath216 -\mu_1 \sum_{i } a_{1i } \,p_{1i } \;,\end{aligned}\ ] ] where @xmath210 is the lagrange multiplier . since @xmath217 , eq .
( [ eq15 ] ) represents a convex optimization problem .
the optimal @xmath218 with diagonal elements given by ( [ lin30 ] ) is simply obtained by applying the kkt conditions to ( [ eq15 ] ) .
[ lin_cor1 ] for any @xmath192 , @xmath193 , let @xmath219 and @xmath220 represent the corner points given by ( [ lin27 ] ) for the optimal @xmath218 and @xmath221 , given by ( [ lin30 ] ) and ( [ lin31 ] ) .
the achievable secrecy rate region of the above approach under the average power constraint ( [ lin2 ] ) is the convex hull of all obtained corner points and is given by @xmath222 it is interesting to note that , unlike the conventional broadcast channel without secrecy constraints where uniform power allocation is optimal in maximizing the sum - rate in the high snr regime @xcite , the high snr power allocation for the bc with confidential messages is a special form of waterfilling as described in the following lemma .
[ lin_lemext1 ] for high snr @xmath223 , the asymptotic optimal power allocations given by ( [ lin30])-([lin31 ] ) are @xmath224 to show ( [ linext1 ] ) we note that @xmath225 when @xmath226 . thus ( [ lin30 ] ) can be written as @xmath227 ( [ linext2 ] ) is proved similarly .
it is also worth noting that the solution in ( [ lin30])-([lin31 ] ) approaches the standard point - to - point mimo waterfilling solution when one of the channels is dominant . for example , let @xmath228 .
we will show that the optimal input covariance simplifies to the waterfilling solution for @xmath229 , given by @xmath230 , where unitary @xmath231 and diagonal @xmath232 are obtained from the eigenvalue decomposition @xmath233 .
the capacity of the point - to - point mimo gaussian link is @xmath234 when @xmath235 , we note from ( [ lin20 ] ) and ( [ lin21 ] ) that @xmath236 , @xmath237 , @xmath238 , and @xmath239 .
consequently , @xmath240 and @xmath241 , where @xmath242 is a diagonal matrix with diagonal elements given by ( [ lin30 ] ) .
the average power constraint in ( [ lin29 ] ) becomes @xmath243 , where @xmath244 when @xmath235 .
thus the @xmath202 diagonal element of @xmath245 converges to the @xmath202 diagonal element of @xmath246 . starting from ( [ lin30 ] ) and applying lhpital s rule , when @xmath228 and hence @xmath247 , we have @xmath248 , and consequently , @xmath249 here we focus on the bc in ( [ lin1 ] ) for the miso case under an average power constraint , where both receivers have a single antenna and the transmitter has @xmath250 antennas : @xmath251 where the channels are represented by the @xmath252 vectors @xmath253 and @xmath254 .
the miso case is the only bc scenario whose secrecy capacity region under the _ average _ power constraint ( [ lin2 ] ) is characterized in closed - form . in particular , it was shown in @xcite that @xmath255 where @xmath256 is the secrecy rate pair on the pareto boundary of the secrecy capacity region for the power fraction @xmath192 , @xmath193 , where power @xmath257 is allocated to receiver 1 s message and @xmath258 is allocated to receiver 2 s message .
furthermore , we have @xcite @xmath259 where @xmath260 @xmath261 is the unit length principal generalized eigenvector of @xmath262 , @xmath263 is the largest generalized eigenvalue of latexmath:[\[(\bi+\frac{(1-\alpha)p_t}{1+\alpha p_t length generalized eigenvector corresponding to @xmath263 .
note that the achievablity of ( [ lin34 ] ) is still based on s - dpc .
while we could have just used the results of section [ sec : mimo ] for the miso case , we will see that the advantage of considering a different approach here is that we obtain a more succinct expression for the achievable secrecy rate region for linear precoding , and we are able to quantify the loss in secrecy rate incurred by linear precoding under the average power constraint compared with @xmath265 .
this was not possible in the mimo case . referring to ( [ lin6 ] ) ,
it was shown in @xcite that for the secrecy rate pair given by ( [ lin34 ] ) , @xmath40 and @xmath41 have covariance matrices @xmath266 and @xmath267 , respectively .
thus , the specific input covariance matrix that attains ( [ lin34 ] ) is given by @xmath268 where @xmath269 and @xmath270 .
equivalently , one can say that under the _ matrix _ power constraint @xmath271 , the corner point of the corresponding rectangular secrecy capacity region is given by ( [ lin34 ] ) .
the union of these corner points constructs the pareto boundary of the secrecy capacity region under the _ average _ power constraint , where any point on the boundary is given by ( [ lin34 ] ) for a different @xmath192 and is achieved under the matrix power constraint @xmath271 given by ( [ lin35 ] ) . using the above fact
, we now present a different linear precoding scheme as an alternative to corollary [ lin_cor1 ] for the miso bc under the average power constraint ( [ lin2 ] ) .
[ lin_cor2 ] using the linear precoding scheme proposed in theorem [ lin_lem2 ] for the miso bc under an average power constraint , the following secrecy rate region is achievable : @xmath272 where @xmath273 @xmath274 and @xmath275 are given by ( [ lin34 ] ) , @xmath276 and where @xmath277 is the unit length principal generalized eigenvector of @xmath278 . from remark [ lin_remetx1 ] , and by noting that for any miso bc
, @xmath279 has at most 2 non - zero eigenvalues , any miso bc can be modeled with a scenario involving just two transmit antennas . thus , without loss of generality , we assume that @xmath280 . from theorem [ lin_lem2 ] , we only need to characterize @xmath281 and @xmath282 , where @xmath281 ( @xmath282 ) is the generalized eigenvector of the pencil @xmath283 corresponding to the generalized eigenvalue larger ( less ) than 1 , @xmath284 ( @xmath285 ) . from ( [ lin6 ] ) and ( [ lin8 ] ) , the covariance matrix of @xmath40 can be rewritten as @xmath286\left [ \begin{array}{ccc } ( \bc_1^h\bc_1)^{-1 } & 0\\ 0 & 0 \end{array } \right][\bc_1\ ; \bc_2]^h \bs_q^{\frac{1}{2 } } = \frac{1}{\bc_1^h\bc_1}\,\bs_q^{\frac{1}{2 } } \bc_1\;\bc_1^h \bs_q^{\frac{1}{2 } } \ ; .\end{aligned}\ ] ] comparing ( [ lin38 ] ) with the covariance matrix of @xmath40 reported in @xcite , we have @xmath287 .
this results in is required for a precise equaltiy , but since this term disappears in the final result , we simply ignore it . ]
@xmath288 on the other hand , from the definition of @xmath281 and @xmath282 ( see ( [ lin10])-([lin11 ] ) for example ) , we have @xmath289^h\left[\bs_q^{\frac{1}{2}}\bh\bh^h\bs_q^{\frac{1}{2}}+\bi\right][\bc_1\ ; \bc_2]=\left[\begin{array}{ccc}\lambda_1 & 0\\ 0 & \lambda_2\end{array}\right ] = \left[\begin{array}{ccc}\gamma_1(\alpha ) & 0\\ 0 & \gamma^{-1}_2(\alpha)\end{array}\right]\\ & [ \bc_1\ ; \bc_2]^h\left[\bs_q^{\frac{1}{2}}\bg\bg^h\bs_q^{\frac{1}{2}}+\bi\right][\bc_1\ ; \bc_2]=\bi \end{split}\end{aligned}\ ] ] , @xmath290.,width=288,height=288 ] where @xmath291 and @xmath263 are defined after ( [ lin34 ] ) , and the fact that @xmath292 and @xmath293 comes from the argument after ( [ lin35 ] ) and by comparing ( [ lin13 ] ) and ( [ lin34 ] ) . substituting ( [ lin39 ] ) in ( [ lin40 ] ) , after some simple calculations , @xmath281 can be explicitly written as in ( [ lin41 ] ) . recalling that @xmath281 is the principal generalized eigenvector of ( [ lin37 ] ) and @xmath282 , which corresponds to the smallest generalized eigenvalue of the pencil ( [ lin37 ] ) , is the principal generalized eigenvector of the pencil @xmath294 we obtain ( [ lin42 ] ) . the proof is completed by using ( [ lin41 ] ) and ( [ lin42 ] ) in ( [ lin19 ] ) .
, @xmath295.,width=288,height=288 ] in this section , we provide numerical examples to illustrate the achievable secrecy rate region of the mimo gaussian bc under the average power constraint ( [ lin2 ] ) . in the first example
, we have @xmath296 , @xmath297 $ ] and @xmath298 $ ] , which is identical to the case studied in ( * ? ? ?
3 ( d ) ) . fig .
[ lin_exm1 ] compares the achievable secrecy rate region of the proposed linear precoding scheme in section iv - a with the secrecy capacity region obtained by the optimal s - dpc approach together with an exhaustive search over suitable matrix constraints , as described in section ii .
we see that in this example , the performance of the proposed linear precoding approach is essentially identical to that of the optimal s - dpc scheme . in the next example
, we study the miso bc for @xmath299 . fig .
[ lin_exm41 ] shows the average secrecy rate regions for s - dpc and the suboptimal linear precoding algorithms described in corollary 1 and 2 .
this plot is based on an average of over 30000 channel realizations , where the channel coefficients were generated as independent @xmath300 random variables .
we see that corollary 2 provides near optimal performance when @xmath301 or @xmath302 , while corollary 1 is better for in - between values of @xmath192 .
the degradation of using linear precoding with corollary 1 is never above 15% for any @xmath192 .
we have shown that for a two - user gaussian bc with an arbitrary number of antennas at each node , when the channel input is under the matrix power constraint , linear precoding is optimal and achieves the secrecy capacity region attained by the optimal s - dpc approach if the matrix constraint satisfies a specific condition .
we characterized the form of the linear precoding that achieves the secrecy capacity region in such cases , and we quantified the maximum loss in secrecy rate that occurs if the matrix power constraint does not satisfy the given condition . based on these observations
, we then formulated a sub - optimal approach for the general mimo scenario based on linear precoding for the case of an average power constraint , for which no known characterization of the secrecy capacity region exists .
we also studied the miso case in detail .
numerical results indicate that the proposed linear precoding approaches yield secrecy rate regions that are close to the secrecy capacity achieved by s - dpc .
from ( [ lin17 ] ) , we have @xmath303 the covariance @xmath42 , given by ( [ lin8 ] ) , can be rewritten as @xmath304 \left[\begin{array}{ccc}(\bc_1^h\bc_1)^{-1 } & \b0\\ \b0 & \b0\end{array}\right]\left[\begin{array}{ccc}\bc^h_1\\\bc^h_2\end{array}\right ] \bs^{\frac{1}{2 } } \nonumber\\ & = \bs^{\frac{1}{2}}\,\bc_1 ( \bc_1^h\bc_1)^{-1 } \bc^h_1\ , \bs^{\frac{1}{2 } } = \bs^{\frac{1}{2}}\,\bp_{\bc_1}\,\bs^{\frac{1}{2 } } \ ; , \end{aligned}\ ] ] where @xmath305 is the projection matrix onto the column space of @xmath61 .
moreover , let @xmath306 be the projection onto the space orthogonal to @xmath61 .
consequently , we have @xmath307\bc^h \bs^{\frac{1}{2 } } \ ; , \label{linap7}\end{aligned}\ ] ] where in ( [ linap6 ] ) , @xmath308 comes from the fact that @xmath70 , and @xmath309 $ ] is full - rank .
following the same steps as in the proof of ( * ? ? ?
* lemma 2 ) or ( * ? ? ?
b ) , we can convert the case when @xmath69 , @xmath310 , to the case where @xmath311 with the same secrecy capacity region . from ( [ lin10 ] ) and ( [ lin11 ] ) we have @xmath312\bc^{-1}-\bi\right ] \bs^{-1/2 } \\ & \bg^h\bg= \bs^{-1/2}\left[\bc^{-h}\bc^{-1}-\bi\right ] \bs^{-1/2 } \ ; . \end{split}\end{aligned}\ ] ] using ( [ linap7 ] ) and ( [ linap8 ] ) , we have : @xmath313 \bc^h\cdot\left[\bc^{-h}\left[\begin{array}{ccc}\mathbf{\lambda}_1 & 0\\ 0 & \mathbf{\lambda}_2\end{array}\right]\bc^{-1}-\bi\right ] \bs^{-1/2}\right|\nonumber \\ & = \left|\bi+ \left[\begin{array}{ccc } \b0 & \b0\\\b0 & ( \bc_2^h\bc_2)^{-1}\end{array}\right ] \cdot\left[\left[\begin{array}{ccc}\mathbf{\lambda}_1 & 0\\ 0 & \mathbf{\lambda}_2\end{array}\right]-\bc^h\bc\right ] \right| \label{linap9 } \\ & = \left|\left[\begin{array}{ccc}\bi & \b0\\\b0 & ( \bc_2^h\bc_2)^{-1}\mathbf{\lambda}_2\end{array } \right]\right| \label{linap10 } \\ & = \left|(\bc_2^h\bc_2)^{-1}\mathbf{\lambda}_2\right| = \left|(\bc_2^h\bc_2)^{-1}\right|\cdot\left|\mathbf{\lambda}_2\right| \ ; , \label{linap11}\end{aligned}\ ] ] where ( [ linap9 ] ) comes from the fact that @xmath314 .
finally , ( [ linap10 ] ) holds since @xmath71 and @xmath315 is block diagonal .
similarly , one can show that @xmath316 and @xmath317\bc^{-1}\right| = \left|(\bc^h\bc)^{-1}\right|\cdot\left|\mathbf{\lambda}_1\right| \cdot\left|\mathbf{\lambda}_2\right| \nonumber\\ & = \left|(\bc_1^h\bc_1)^{-1}\right|\cdot\left|(\bc_2^h\bc_2)^{-1}\right|\cdot\left|\mathbf{\lambda}_1\right| \cdot\left|\mathbf{\lambda}_2\right| \;.
\label{linap14}\end{aligned}\ ] ] substituting ( [ linap11 ] ) , ( [ linap13 ] ) and ( [ linap14 ] ) in ( [ linap4 ] ) , we have @xmath318 , and this completes the proof .
from ( [ lin10])-([lin11 ] ) , we know that @xmath319 , where @xmath56 represents number of generalized eigenvalues of the pencil ( [ lin9 ] ) that are greater than 1 . from ( [ lin10])-([lin11 ] ) ,
we have @xmath320\bc_1=\mathbf{\lambda}_1 \label{linap1 } \\ & \bc_1^h\left[\bs^{\frac{1}{2}}\bg^h\bg\bs^{\frac{1}{2}}+\bi\right]\bc_1=\bi \;.\label{linap2}\end{aligned}\ ] ] subtracting ( [ linap1 ] ) from ( [ linap2 ] ) , a straightforward computation yields @xmath321\bs^{\frac{1}{2}}\bc_1= \mathbf{\lambda}_1-\bi \succ\b0 \;.\end{aligned}\ ] ] from ( [ linap3 ] ) , we have @xmath322\bs^{\frac{1}{2}}\bc_1\succ\b0 $ ] , from which it follows that @xmath323 . similarly one can show that @xmath324 , where @xmath325 corresponds to the generalized eigenvalues of the pencil ( [ lin9 ] ) which are less than 1 , and @xmath326 represents number of negative eigenvalues of @xmath82 .
we want to show that @xmath122 and @xmath123 commute , where @xmath389 let the invertible matrix @xmath390 and diagonal matrix @xmath391 respectively represent the generalized eigenvectors and eigenvalues of @xmath392 , so that @xmath393\widehat{\bc}=\widehat{\mathbf{\lambda } } \label{linap26}\\ & \widehat{\bc}^h\left[\bw\bg^h\bg\bw+\bi\right]\widehat{\bc}=\bi \;. \label{linap27}\end{aligned}\ ] ] adding ( [ linap26 ] ) and ( [ linap27 ] ) , we have @xmath394\widehat{\bc}= 3\,\widehat{\bc}^h\widehat{\bc}=(\widehat{\mathbf{\lambda}}+\bi ) \;,\end{aligned}\ ] ] from which it results that @xmath390 must be of the form @xcite @xmath395 where @xmath125 is an unknown unitary matrix . in the following , as we continue the proof , @xmath125 is characterized too . substituting ( [ linap28 ] ) in ( [ linap26 ] ) and ( [ linap27 ] ) , it is revealed that the unitary matrix @xmath125 represents the common set of eigenvectors for the matrices @xmath396 and @xmath397 , and thus both matrices commute . in particular , @xmath398\bphi_{\bw } & = 3\,\widehat{\mathbf{\lambda } } ( \widehat{\mathbf{\lambda}}+\bi)^{-1}=3\,(\widehat{\mathbf{\lambda}}^{-1}+\bi)^{-1 } \\
\bphi_{\bw}^h\left[\bw\bg^h\bg\bw+\bi\right]\bphi_{\bw}&=3\ , ( \widehat{\mathbf{\lambda}}+\bi)^{-1}\;. \end{split}\end{aligned}\ ] ] consequently , @xmath128 and @xmath129 are diagonal : @xmath399 it is interesting to note that , since @xmath400 and @xmath401 , we have @xmath402 .
we first consider the generalized eigenvalue decomposition for @xmath403 where @xmath164 is given by ( [ lin23 ] ) and @xmath404\overline{\bc}_{\bw}=\mathbf{\lambda}_{\bw}\\ & \overline{\bc}_{\bw}^h\left[\bt_{\bw}^h\bg^h\bg\bt_{\bw}+\bi\right]\overline{\bc}_{\bw}=\bi \ ; . \end{split}\end{aligned}\ ] ] using ( [ lin21 ] ) , and noting that @xmath125 is unitary and @xmath151 is diagonal , a straightforward calculation yields @xmath405\overline{\bc}_{\bw}=\mathbf{\lambda}_{\bw}\\ & \overline{\bc}_{\bw}^h\left[\bsig_2\bp+\bi\right]\overline{\bc}_{\bw}=\bi\ ; , \end{split}\end{aligned}\ ] ] where @xmath128 and @xmath129 are respectively ( diagonal ) eigenvalue matrices of @xmath122 and @xmath123 , as given by ( [ lin21 ] ) .
thus , @xmath406 is diagonal and is given by @xmath407 consequently , we have @xmath408 .
let @xmath409 , @xmath410 and @xmath411 represent the @xmath202 diagonal elements of @xmath128 , @xmath129 and @xmath151 , respectively .
we note that for any @xmath411 , @xmath412 iff @xmath413 .
thus , based on the argument that we made after lemma [ lin_lem3 ] , the first @xmath127 diagonal elements of @xmath414 represent generalized eigenvalues greater than 1 . letting @xmath415\end{aligned}\ ] ] where @xmath157 is @xmath416 and @xmath158 is @xmath417
, we have : @xmath418 = \left[\begin{array}{ccc}\mathbf{\lambda}_{1\bw } & \b0\\\b0 & \mathbf{\lambda}_{2\bw}\end{array } \right ] \ ; , \end{aligned}\ ] ] where @xmath133 and @xmath135 ( @xmath111 ) are given by ( [ lin22 ] ) .
consequently , ( [ linap32 ] ) can be rewritten as @xmath419= \left[\begin{array}{ccc}\left(\bsig_{2\rho}\bp_1+\bi\right)^{-\frac{1}{2 } } & \b0\\\b0 & \left(\bsig_{2\bar{\rho}}\bp_2+\bi\right)^{-\frac{1}{2}}\end{array}\right ] \;.\end{aligned}\ ] ] @xmath141 , linear precoding is an optimal solution for the bc under the matrix power constraint @xmath420 , where @xmath164 is given by ( [ lin23 ] ) .
more precisely , from theorem [ lin_thm1 ] , @xmath72 is optimal , where @xmath33 and @xmath34 are independent gaussian precoders , respectively corresponding to @xmath15 and @xmath16 with zero means and covariance matrices @xmath421 and @xmath422 , where @xmath421 is given by @xmath423\bc_{\bw}^h \bs_{\bw}^{\frac{1}{2}}\end{aligned}\ ] ] and @xmath424 is the generalized eigenvector matrix for @xmath425 we note that there exists a unitary matrix @xmath331 for which @xmath426 @xcite , where @xmath427 .
we also note that , from remark [ lin_remap2 ] , @xmath428 and @xmath429 .
thus , @xmath421 can be rewritten as @xmath430\overline{\bc}_{\bw}^h \bt_{\bw}^h = \bt_{\bw}\left[\begin{array}{ccc}\bi & \b0\\\b0 & \b0\end{array } \right]\bt_{\bw}^h \label{linap39 } \\ & = \bw\bphi_{\bw}\bp^{\frac{1}{2}}\left[\begin{array}{ccc}\bi & \b0\\\b0 & \b0\end{array } \right]\bp^{\frac{1}{2}}\bphi^h_{\bw}\bw = \bw\bphi_{\bw}\left[\begin{array}{ccc}\bp_1 & \b0\\\b0 & \b0\end{array } \right]\bphi^h_{\bw}\bw \ ; , \label{linap40}\end{aligned}\ ] ] where ( [ linap39 ] ) comes from ( [ linap36 ] ) , and ( [ linap40 ] ) comes from ( [ linap34 ] ) .
consequently , @xmath422 can be written as @xmath431\bphi^h_{\bw}\bw \ ; . \label{linap41}\end{aligned}\ ] ] from ( [ linap40 ] ) and ( [ linap41 ] ) , under the matrix power constraint @xmath150 given by ( [ lin24 ] ) , the optimal linear precoding is @xmath72 , where precoding signals @xmath33 and @xmath34 are independent gaussian vectors with zero means and covariance matrices given by ( [ linap40 ] ) and ( [ linap41 ] ) , respectively .
alternatively , the optimal precoder can be represented as @xmath432 $ ] , where precoding signals @xmath433 and @xmath434 are independent gaussian vectors with zero means and diagonal covariance matrices respectively given by @xmath157 and @xmath158 . in both cases @xmath435 , and the same secrecy rate region is achieved . 1 a. wyner , `` the wire - tap channel , '' _ bell .
j. _ , vol .
54 , no . 8 , pp . 1355 - 1387 , jan . 1975 .
r. liu , i. maric , p. spasojevic , and r. d. yates , discrete memoryless interference and broadcast channels with confidential messages : secrecy rate regions , " _ ieee trans .
inf . theory _
54 , no . 6 , pp . 2493 - 2512 , june 2008 .
r. liu and h. v. poor , secrecy capacity region of a multiple - antenna gaussian broadcast channel with confidential messages , " _ ieee trans .
inf . theory _
1235 - 1249 , mar .
r. liu , t. liu , h. v. poor , and s. shamai , multiple - input multiple - output gaussian broadcast channels with confidential messages , " _ ieee trans .
inf . theory _ ,
4215 - 4227 , 2010 .
q. h. spencer , a. l. swindlehurst , and m. haardt , zero - forcing methods for downlink spatial multiplexing in multiuser mimo channels , " _ ieee trans .
signal processing _ , vol .
461 - 471 , feb . 2004 .
t. yoo and a. goldsmith , on the optimality of multi - antenna broadcast scheduling using zero - forcing beamforming , " _ ieee j. select .
areas commun .
_ , special issue on 4 g wireless systems , vol .
3 , pp.528 - 541 , mar .
a. wiesel , y. eldar , and s. shamai , linear precoding via conic optimization for fixed mimo receivers , " _ ieee trans .
signal processing _ , vol .
161 - 176 , jan . 2006 .
a. fakoorian and a. l. swindlehurst , `` dirty paper coding versus linear gsvd - based precoding in mimo broadcast channel with confidential messages , '' in _ proc .
ieee globecom _ , dec
a. khisti and g. wornell , `` secure transmission with multiple antennas ii : the mimome wiretap channel , '' _ ieee trans .
56 , no . 11 , pp . 5515 - 5532 , 2010 .
a. fakoorian and a. l. swindlehurst , `` optimal power allocation for the gsvd based mimo gaussian wiretap channel , '' in _ isit _ , july 2012 .
r. bustin , r. liu , h. v. poor , and s. shamai ( shitz ) , `` a mmse approach to the secrecy capacity of the mimo gaussian wiretap channel , '' _ eurasip journal on wireless comm . and net .
2009 , article i d 370970 , 8 pages , 2009 .
r. a. horn and c. r. johnson , _ matrix analysis _ , university press , cambridge , uk , 1985 .
h. weingarten , y. steinberg , and s. shamai ( shitz ) , `` the capacity region of the gaussian multiple - input multiple - output broadcast channel , '' _ ieee trans . inf .
9 , pp . 3936 - 3964 , 2006 .
d. nion , `` a tensor framework for nonunitary joint block diagonalization , '' _ ieee trans .
signal processing _ , vol .
4585 - 4594 , oct .
s. a. a. fakoorian and a. l. swindlehurst , `` mimo interference channel with confidential messages : achievable secrecy rates and beamforming design , '' _ ieee trans . on inf .
forensics and security _ , vol .
j. lee , and n. jindal , `` high snr analysis for mimo broadcast channels : dirty paper coding versus linear precoding , '' _ ieee trans .
inf . theory _ ,
4787 - 4792 , dec .
we want to characterize the matrices @xmath69 for which @xmath327 has generalized eigenvectors with orthogonal @xmath61 and @xmath63 . for any positive semidefinite matrix @xmath328
, there exists a matrix @xmath329 such that @xmath84 @xcite .
more precisely , @xmath330 , where @xmath331 can be any @xmath332 unitary matrix ; thus @xmath85 is not unique .
[ lin_remap2 ] let the invertible matrix @xmath333 and the diagonal matrix @xmath334 respectively represent the generalized eigenvectors and eigenvalues of @xmath335 so that @xmath336\overline{\bc}=\overline{\mathbf{\lambda}}\\ & \overline{\bc}^h\left[\bt^h\bg^h\bg\bt+\bi\right]\overline{\bc}=\bi \ ; , \end{split}\end{aligned}\ ] ] where @xmath330 for a given unitary matrix @xmath331 . by comparing ( [ lin10 ] ) and ( [ linap_ex1 ] ) , one can confirm that @xmath337 and @xmath338 , where @xmath339 and @xmath95 are respectively the generalized eigenvectors and eigenvalues of ( [ linap22 ] ) , as given by ( [ lin10 ] ) .
also note that , for any unitary @xmath331 , @xmath340 .
thus , finding a @xmath69 such that ( [ linap22 ] ) has orthogonal @xmath61 and @xmath63 ( block diagonal @xmath315 ) is equivalent to finding a @xmath85 , @xmath84 , such that ( [ linap23 ] ) has orthogonal @xmath341 and @xmath342 ( block diagonal @xmath343 ) .
the _ if _ part of theorem [ lin_thm2 ] is easy to show .
we want to show that if @xmath84 and @xmath85 simultaneously block diagonalizes @xmath86 and @xmath87 , as given by ( [ linshrink1 ] ) such that @xmath344 and @xmath345 , then @xmath71 . from the definition of the generalized eigenvalue decomposition , we have @xmath346\overline{\bc}= \overline{\bc}^h \left[\begin{array}{ccc}\bi+\bk_{\bh1 } & \b0\\ \b0 & \bi+\bk_{\bh2}\end{array}\right]\overline{\bc}= \left[\begin{array}{ccc}\bd_1 & \b0\\ \b0 & \bd_2 \end{array}\right]\\ & \overline{\bc}^h\left[\bt^h\bg^h\bg\bt+\bi\right]\overline{\bc}=\overline{\bc}^h \left[\begin{array}{ccc}\bi+\bk_{\bg1 } & \b0\\ \b0 & \bi+\bk_{\bg2}\end{array}\right]\overline{\bc}= \left[\begin{array}{ccc}\bi & \b0\\ \b0 & \bi \end{array}\right ] \ ; , \\ \end{split}\end{aligned}\ ] ] from which we have @xmath347=\left[\begin{array}{ccc}\overline{\bc}_{11 } & \b0\\ \b0 & \overline{\bc}_{22}\end{array}\right ] \;,\ ] ] where the invertible matrix @xmath348 and diagonal matrix @xmath349 are respectively the generalized eigenvectors and eigenvalues of @xmath350 . since @xmath351 , then @xmath352 , which shows that @xmath341 corresponds to generalized eigenvalues that are bigger than or equal to one .
we have a similar definition for @xmath353 and diagonal matrix @xmath354 , corresponding to @xmath355 . finally , since @xmath343 is block diagonal , then @xmath315 , where @xmath339 is the generalized eigenvector matrix of ( [ linap22 ] ) , is block diagonal as well .
this completes the _ if _ part of the theorem . in the following , we prove the _ only if _ part of theorem [ lin_thm2 ] ; _ i.e. , _ we show that if @xmath69 results in ( [ linap22 ] ) having orthogonal @xmath61 and @xmath63 , then there must exist a square matrix @xmath85 such that @xmath84 and @xmath356 and @xmath357 are simultaneously block diagonalized as in ( [ linshrink1 ] ) with @xmath344 and @xmath345 .
let @xmath358 have the eigenvalue decomposition @xmath359 , where @xmath360 is unitary and @xmath361 is a positive semidefinite diagonal matrix .
also let @xmath362 have the eigenvalue decomposition @xmath363 , where @xmath364 is unitary and @xmath365 is a positive definite diagonal matrix .
one can easily confirm that @xcite @xmath366 and @xmath367 , where @xmath339 and @xmath95 are respectively the generalized eigenvectors and eigenvalues of ( [ linap22 ] ) . also let @xmath339 be ordered such that @xmath309 $ ] , where @xmath61 corresponds to the generalized eigenvalues bigger than ( or equal to ) 1 .
we have @xmath368 from ( [ linap25 ] ) , @xmath315 is block diagonal iff the unitary matrix @xmath364 is block diagonal . recalling that @xmath364 is the eigenvector matrix of @xmath362 , a block diagonal @xmath364 leads to @xmath362 , and consequently @xmath369 must be block diagonal .
thus , if @xmath315 is block diagonal , _
@xmath71 , there must exist a unitary matrix @xmath360 such that @xmath369 and @xmath370 are simultaneously block diagonal .
is actually diagonal , and hence also block diagonal . ] letting @xmath371 results in ( [ linap24 ] ) , for which we must have @xmath344 and @xmath345 , otherwise it contradicts the ordering of @xmath309 $ ] .
this completes the proof .
[ lin_remap1 ] by applying the schur complement lemma @xcite on @xmath372^h\left[\bc_1\quad
\bc_2\right]=\left [ \begin{array}{ccc}\bc_1^h\bc_1 & \bc_1^h\bc_2\\ \bc_2^h\bc_1 & \bc_2^h\bc_2\end{array}\right]\ ] ] and recalling the fact that @xmath339 is full - rank , we have that @xmath373 is full rank .
similarly , one can show that @xmath374 exists . also , we have @xmath375 @xmath376 .
define @xmath377 $ ] , so that @xmath378^h\left[\bp^\perp_{\bc_2}\bc_1\quad \bc_2\right]=\left [ \begin{array}{ccc}\bc_1^h\bp^\perp_{\bc_2}\bc_1 & \b0\\ \b0 & \bc_2^h\bc_2\end{array}\right]\;.\ ] ] consequently , we can write @xmath379\widehat{\bc}^h\end{aligned}\ ] ] and @xmath380\widehat{\bc}^h \ ; .\end{aligned}\ ] ] in the following we show the achievablity of @xmath77 in ( [ lin19 ] ) .
the achievablity of @xmath381 is obtained in a similar manner .
since @xmath33 and @xmath34 in theorem [ lin_lem2 ] are independent , from ( [ lin17 ] ) we have @xmath382 recalling ( [ linap8 ] ) , we have @xmath383 where we used remark [ lin_remap1 ] to obtain ( [ linap18 ] ) . from ( [ linap15 ] ) , we have @xmath384 \widehat{\bc}^h\cdot\left[\bc^{-h}\left[\begin{array}{ccc}\mathbf{\lambda}_1 & 0\\ 0 & \mathbf{\lambda}_2\end{array}\right]\bc^{-1}-\bi\right ] \right|\nonumber \\ & = \left|\bi+\left[\begin{array}{ccc } \b0 & \b0\\\b0 & ( \bc_2^h\bc_2)^{-1}\end{array}\right ] \cdot\left[\widehat{\bc}^h\bc^{-h}\left[\begin{array}{ccc}\mathbf{\lambda}_1 & 0\\ 0 & \mathbf{\lambda}_2\end{array}\right]\bc^{-1}\widehat{\bc}-\widehat{\bc}^h\widehat{\bc } \right ] \right|\nonumber \\ & = \left|\bi+\left[\begin{array}{ccc } \b0 & \b0\\\b0 & ( \bc_2^h\bc_2)^{-1}\end{array}\right ] \cdot\left[\left[\begin{array}{ccc}\bi & \bn^h\\ \b0 & \bi\end{array}\right ] \left[\begin{array}{ccc}\mathbf{\lambda}_1 & 0\\ 0 & \mathbf{\lambda}_2\end{array}\right]\left[\begin{array}{ccc}\bi & \b0\\ \bn & \bi\end{array}\right ] -\widehat{\bc}^h\widehat{\bc } \right ] \right| \label{linap19}\\ & = \left|\bi+\left[\begin{array}{ccc } \b0 & \b0\\\b0 & ( \bc_2^h\bc_2)^{-1}\end{array}\right ] \cdot\left[\left[\begin{array}{ccc}\mathbf{\lambda}_1+\bn^h\mathbf{\lambda}_2\bn & \bn^h\mathbf{\lambda}_2\\ \mathbf{\lambda}_2\bn & \mathbf{\lambda}_2\end{array}\right ] -\widehat{\bc}^h\widehat{\bc } \right ] \right| \nonumber\\ & = \left|\left[\begin{array}{ccc } \bi & \b0 \\ ( \bc_2^h\bc_2)^{-1}\mathbf{\lambda}_2\bn & ( \bc_2^h\bc_2)^{-1}\mathbf{\lambda}_2\end{array}\right ] \right| \nonumber\\ & = \left|(\bc_2^h\bc_2)^{-1}\mathbf{\lambda}_2\right| = \left|(\bc_2^h\bc_2)^{-1}\right|\cdot\left|\mathbf{\lambda}_2\right| \ ; , \label{linap20}\end{aligned}\ ] ] where in ( [ linap19 ] ) , @xmath385 , and we used the fact that @xmath386\ , \left[\bp^\perp_{\bc_2}\bc_1 \quad \bc_2\right ] = \left[\begin{array}{ccc}\bi & \b0\\ \bn & \bi\end{array}\right ] \;.\end{aligned}\ ] ] similarly , we have @xmath387 \cdot\left[\left[\begin{array}{ccc}\bi & \bn^h\\ \b0 & \bi\end{array}\right ] \left[\begin{array}{ccc}\bi & \b0\\ \bn & \bi\end{array}\right ] -\widehat{\bc}^h\widehat{\bc } \right ] \right| \nonumber\\ & = \left|\bi+\left[\begin{array}{ccc } ( \bc_1^h\bp^\perp_{\bc_2}\bc_1)^{-1 } & \b0\\\b0 & \b0\end{array}\right ] \cdot\left[\left[\begin{array}{ccc}\bi+\bn^h\bn & \bn^h\\ \bn & \bi\end{array}\right ] -\widehat{\bc}^h\widehat{\bc } \right ] \right| \nonumber\\ & = \left|\left[\begin{array}{ccc } ( \bc_1^h\bp^\perp_{\bc_2}\bc_1)^{-1 } ( \bi+\bn^h\bn ) & ( \bc_1^h\bp^\perp_{\bc_2}\bc_1)^{-1}\bn^h \\ \b0 & \bi\end{array}\right ] \right| \nonumber\\ & = \left|(\bc_1^h\bp^\perp_{\bc_2}\bc_1)^{-1}\right|\cdot\left|\bi+\bn^h\bn\right| \;. \label{linap21}\end{aligned}\ ] ] subsituting ( [ linap18 ] ) , ( [ linap20 ] ) and ( [ linap21 ] ) in ( [ linap17 ] ) , we have @xmath388 , which completes the proof . | we study the optimality of linear precoding for the two - receiver multiple - input multiple - output ( mimo ) gaussian broadcast channel ( bc ) with confidential messages .
secret dirty - paper coding ( s - dpc ) is optimal under an input covariance constraint , but there is no computable secrecy capacity expression for the general mimo case under an average power constraint . in principle , for this case , the secrecy capacity region could be found through an exhaustive search over the set of all possible matrix power constraints . clearly , this search , coupled with the complexity of dirty - paper encoding and decoding , motivates the consideration of low complexity linear precoding as an alternative .
we prove that for a two - user mimo gaussian bc under an input covariance constraint , linear precoding is optimal and achieves the same secrecy rate region as s - dpc if the input covariance constraint satisfies a specific condition , and we characterize the corresponding optimal linear precoders .
we then use this result to derive a closed - form sub - optimal algorithm based on linear precoding for an average power constraint .
numerical results indicate that the secrecy rate region achieved by this algorithm is close to that obtained by the optimal s - dpc approach with a search over all suitable input covariance matrices . |
the ricci flow has been introduced by r. hamilton @xcite with the goal of providing an analytic approach to thurston s geometrization conjecture for three - manifolds @xcite@xmath0 . inspired by the theory of harmonic maps , he considered the geometric evolution equation obtained when one evolves a riemannian metric @xmath1 , on a three - manifold @xmath2 , in the direction of its ricci tensor@xcite @xmath3 , _
i.e. _ @xmath4 in recent years , this geometric flow has gained extreme popularity thanks to the revolutionary breakthroughs of g. perelman @xcite@xmath0 , who , taking the whole subject by storm , has brought to completion hamilton s approach to thurston s conjecture .
the prominent themes recurring in hamilton s and perelman s works converge to a proof that the ricci flow , coupled to topological surgery , provides a natural technique for factorizing and uniformizing a three - dimensional riemannian manifold @xmath5 into locally homogeneous geometries .
this is a result of vast potential use also in theoretical physics , where the ricci flow often appears in disguise as a natural real - space renormalization group flow .
non - linear @xmath6-model theory , describing quantum strings propagating in a background spacetime , affords the standard case study in such a setting @xmath0 .
another paradigmatical , perhaps even more direct , application occurs in relativistic cosmology @xcite@xmath0 , ( for a series of recent results see also @xcite and the references cited therein ) .
this will be related to the main topic of this talk , and to motivate our interest in it , let us recall that homogeneous and isotropic solutions of einstein s laws of gravity ( the friedman
lemaitre robertson walker ( flrw ) spacetimes ) do not account for inhomogeneities in the universe .
the question whether they do _ on average _ is an issue @xcite that is the subject of considerable debate especially in the recent literature ( see @xcite and follow up references ; comprehensive lists may be found in @xcite and @xcite ) . in any case
, a member of the family of flrw cosmologies ( the so called concordance model that is characterized by a dominating cosmological constant in a spatially flat universe model ) provides a successful _ fitting model _ to a large number of observational data , and the generally held view is that the spatial sections of flrw spacetimes indeed describe the _ physical universe _ on a sufficiently large averaging scale .
this raises an interesting problem in mathematical cosmology : devise a way to explicitly construct a constant curvature metric out of a _ scale dependent _ inhomogeneous distribution of matter and spatial curvature .
it is in such a framework that one makes the basic observation that the ricci flow ( [ mflow ] ) and its linearization , provide a natural technique@xcite for deforming , and under suitable conditions smoothing , the geometrical part of scale dependent cosmological initial data sets .
moreover , by taking advantage of some elementary aspects of perelman s results , this technique also provides a natural and unique way for deforming , along the ricci flow , the matter distribution .
the expectation is that in this way we can define a deformation of cosmological initial data sets into a one - parameter family of initial data whose time evolution , along the evolutive part of einstein s equations , describe the ricci flow deformation of a cosmological spacetime .
to set notation , we emphasize that throughout the paper we shall consider a smooth three - dimensional manifold @xmath2 , which we assume to be closed and without boundary .
we let @xmath7 and @xmath8 be the space of smooth functions and of smooth @xmath9tensor fields on @xmath2 , respectively .
we shall denote by @xmath10 the group of smooth diffeomorphisms of @xmath2 , and by @xmath11 the space of all smooth riemannian metrics over @xmath12 .
the tangent space , @xmath13 , to @xmath11 at @xmath14 can be naturally identified with the space of symmetric bilinear forms @xmath15 over @xmath2 .
the hypothesis of smoothness has been made for simplicity .
results similar to those described below , can be obtained for initial data sets with finite holder or sobolev differentiability .
in such a framework , let us recall that a collection of fields @xmath16 , @xmath17 , @xmath18 , @xmath19 , defined over the three - manifold @xmath2 , characterizes a set @xmath20 , of physical cosmological initial data for einstein equations if and only if the _ matter fields _
@xmath21 verify the weak energy condition @xmath22 , the dominant energy condition @xmath23 , and their coupling with the geometric fields @xmath24 is such as to satisfy the hamiltonian and divergence constraints ; we adopt the summation convention .
the nabla operator denotes covariant derivative with respect to the 3metric .
the units are such that @xmath25 . ] : @xmath26 here @xmath27 is the cosmological constant , @xmath28 , and @xmath29 is the scalar curvature of the riemannian metric @xmath1 . if such a set of admissible data is propagated according to the evolutive part of einstein s equations , then the symmetric tensor field @xmath30 can be interpreted as the extrinsic curvature and @xmath31 as the mean curvature of the embedding @xmath32 of @xmath33 in the spacetime @xmath34 resulting from the evolution of @xmath35 , whereas @xmath36 and @xmath37 are , respectively , identified with the mass density and the momentum density of the material self
gravitating sources on @xmath33 .
the averaging procedure described in @xcite is based on a smooth deformation of the physical initial data @xmath35 into a one - parameter family of initial data sets @xmath38 with @xmath39 being a parameter characterizing the averaging scale .
the general idea is to construct the flow ( [ betadata ] ) in such a way as to represent , as @xmath40 increases , a scale - dependent averaging of @xmath41 , and under suitable hypotheses reducing it to a constant - curvature initial data set @xmath42 where @xmath43 is a constant curvature metric on @xmath2 , @xmath44 is the ( spatially constant ) trace of the extrinsic curvature ( related to the hubble parameter ) , and @xmath45 is the averaged matter density . under the heading of such a general strategy
it is easy to figure out the reasons for an important role played by the ricci flow and its linearization . as we shall recall shortly
, they are natural geometrical flows always defining a non - trivial deformation of the metric @xmath1 and of the extrinsic curvature @xmath30 . moreover , when global , they posses remarkable smoothing properties . for instance , if , for @xmath46 , the scalar curvature @xmath47 of @xmath33 is @xmath48 , and if there exist positive constants @xmath49 , @xmath50 , @xmath51 , not depending on @xmath40 , such that @xmath52,and @xmath53,where @xmath54is the trace
free part of the ricci tensor , then @xcite the solutions @xmath55 of the ( volume normalized ) ricci flow and its linearization exist for all @xmath56 , and the pair @xmath57 , uniformly converges , when @xmath58 , to @xmath59 where @xmath43 is a metric with constant positive sectional curvature , and @xmath60 is some vector field on @xmath2 , possibly depending on @xmath61
. the flow @xmath62 , describing a motion by diffeomorphisms over a constant curvature manifold , can be thought of as representing the smoothing of the geometrical part of an initial data set @xmath35 .
+ it is useful to keep in mind what we can expect and what we can not expect out of such a ricci flow deformation of cosmological initial data set .
let us start by remarking that the family of data ( [ betadata ] ) will correspond to the initial data for physical spacetimes iff the constraints ( [ constraint1 ] ) and ( [ constraints ] ) hold throughout the @xmath40dependent deformation .
this is a very strong requirement and , if we have a technically consistent way @xmath63 of deforming the matter distribution and the geometrical data , then the most natural way of implementing the constraints is to use them to define scale
dependent backreaction fields @xmath64 , @xmath65 describing the non linear interaction between matter averaging and geometrical averaging , _
i.e. _ , @xmath66\;,\label{hamro}\\ \psi _ { a } ( \beta ) & \doteq & j_{a}(\beta ) -(8\pi \,g)^{-1 } \left[\nabla _ { b}k_{\;\,a}^{b}(\beta ) -\nabla _ { a}k(\beta ) \right]\;. \label{divj}\end{aligned}\ ] ] to illustrate how this strategy works , let us concentrate , in this talk , on the characterization of the scalar field @xmath64 , providing the backreaction between matter and geometrical averaging .
the covector field @xmath67 can , in principle , be controlled by the action of a @xmath40dependent diffeomorphism .
however , its analysis requires a subtle interplay with the kinematics of spacetime foliations @xcite@xmath0 , ( _ i.e. _ , how we deal with the lapse function and with the shift vector field in the framework of perelman s approach ) , and will be discussed elsewhere , ( for a pre
perelman approach to this issue see @xcite)@xmath0 .
+ let us start by observing that the matter averaging flow @xmath68 must comply with the preservation of the physical matter content @xmath69 and must be explicitly coupled to the scale of geometrical averaging . in other words ,
if , for some fixed @xmath70 , we consider that part of the matter distribution @xmath71 which is localized in a given region @xmath72 of size @xmath73 , then we should be able to tell from which localized distribution @xmath74 , at @xmath46 , the selected matter content @xmath75 has evolved .
a natural answer to these requirements is provided by perelman s
backward localization@xcite of probability measures on ricci evolving manifolds .
the idea is to probe the ricci flow with a probability measure whose dynamics can localize the regions of the manifold @xmath76 of geometric interest .
this is achieved by considering , along the solution @xmath77 of ( [ mflow ] ) , a @xmath40dependent mapping @xmath78 , in terms of which one constructs on @xmath79 the measure @xmath80 , where @xmath81 is a scale parameter chosen in such a way as to normalize @xmath82 according to the so called _ perelman s coupling_:@xmath83 .
it is easily verified that this is preserved in form along the ricci flow ( [ mflow ] ) , if the mapping @xmath84 and the scale parameter @xmath85 are evolved backward in time @xmath86 according to the coupled flows defined by @xmath87 where @xmath88 is the laplacian with respect to the metric @xmath77 , and @xmath89 , @xmath90 are given ( final ) data . in this connection , note that the equation for @xmath84 is a backward heat equation , and as such the forward evolution @xmath91 is an ill - posed problem .
a direct way for circumventing such a difficulty is to interpret ( [ ourp ] ) according to the following two - steps prescription : ( i ) evolve the metric @xmath92 , say up to some @xmath93 , according to the ricci flow , ( if the flow is global we may let @xmath94 ) ; ( ii ) on the ricci evolved riemannian manifold @xmath95 so obtained , select a function @xmath96 and the corresponding scale parameter @xmath97 , and evolve them , backward in @xmath40 , according to ( [ ourp ] ) .
with these preliminary remarks along the way , let us characterize the various steps involved in constructing the flow ( [ betadata ] ) .
( for ease of exposition , we refer to the standard unnormalized flow ; volume normalization can be enforced by a reparametrization of the deformation parameter ) .
+ ( _ geometrical data deformation _ ) given an initial data set @xmath99 for a cosmological spacetime @xmath100 , the ricci
flow deformation of its geometrical part @xmath101 is defined by the flow @xmath102 , @xmath103 provided by the ( weakly ) parabolic initial value problem ( the ricci flow , proper ) @xmath104<\infty \;,$% \end{tabular } \right .
\label{fflow}\ ] ] and by its linearization , which , by suitably fixing the action of the diffeomorphism group @xmath10 , takes the form of the parabolic initial value problem @xcite @xmath105 where @xmath106 denotes the lichnerowicz
derham laplacian @xmath107 acting on symmetric bilinear forms @xcite@xmath0 .
( _ localization of the deformed data _ ) the geometrical deformation @xmath102 is controlled by the backward localizing flow @xmath108 , @xmath109 , defined by the backward heat kernels for the conjugate @xcite parabolic initial value problem associated with the @xmath110dependent laplace - beltrami and lichnerowicz operators @xmath111 .
the operator @xmath112 , when acting on ( bi)scalars , reduces to @xmath113 .
thus , we can characterize both these kernels in a compact form as solutions of @xmath114 where @xmath115 $ ] , @xmath116 denotes the lichnerowicz
derham laplacian with respect to the variable @xmath117 , the heat kernels @xmath118 and @xmath119 are smooth sections of @xmath120 and @xmath121 , respectively , and finally , @xmath122 is the dirac @xmath123tensorial measure , ( @xmath124 ) on @xmath125 .
the dirac initial condition is understood in the distributional sense , _
i.e. _ , @xmath126 , for any smooth symmetric bilinear form with compact support @xmath127 , and where the limit is meant in the uniform norm on @xmath128 . note that heat kernels for generalized laplacians , such as @xmath129 , ( smoothly ) depending on a one parameter family of metrics @xmath130 , @xmath131 , are briefly dealt with in @xcite@xmath0 .
the delicate setting where the parameter dependence is , as in our case , identified with the parabolic time driving the diffusion of the kernel , is discussed in @xcite@xmath0 , ( see appendix a , 7 for a characterization of the parametrix of the heat kernel in such a case ) , and in @xcite@xmath0
. strictly speaking , in all these works , the analysis is confined to the scalar laplacian , possibly with a potential term , but the theory readily extends to generalized laplacians , always under the assumption that the metric @xmath77 is smooth as @xmath132 .
finally , the kernels for @xmath133 and @xmath134 can both be normalized , along the ricci flow , over the round ( collapsing ) 3sphere @xmath135 .
let @xmath145 be the ricci flow deformation , on @xmath146 $ ] , of the data @xmath147 as defined above .
assume that the underlying ricci flow @xmath148 is of bounded geometry , and let @xmath149 and @xmath118 be the ( backward ) heat kernels ricci - flow conjugated to the laplace beltrami and to the lichnerowicz
derham operator , respectively .
then , for all @xmath150 , @xmath151\,d\mu _ { g(x,\eta ) } \ ; , \label{grepres0}\ ] ] @xmath152 @xmath153 where @xmath156 @xmath157 is the parallel transport operator associated with @xmath158 , @xmath159 is the distance function in @xmath160 , and @xmath161 } ^{ab}_{i'k'}(y , x;\eta ) $ ] are smooth sections @xmath162 , ( depending on the geometry of @xmath158 ) , characterizing the asymptotics of the heat kernel @xmath118 . with an obvious adaptation
, such an asymptotic behavior also extends to @xmath163 and @xmath164 . with these results
it is rather straightforward to provide useful characterizations of the ( intrinsic ) backreaction field @xmath165 . for illustrative purposes ,
let us assume that @xmath166 and @xmath167 , then from ( [ hamro ] ) and the above proposition we get that @xmath168\;.\nonumber\end{aligned}\ ] ] if we further assume that the hamiltonian constraint holds at the _ fixed _ observative scale @xmath139 and that @xmath169 $ ] , ( _ i.e. _ , curvature fluctuates around the _ constant curvature background _
@xmath170 ) , then @xmath171 from which it follows that the intrinsic backreaction field @xmath165 is generated by curvature fluctuations around the given background , as expected .
m.c . would like to dedicate this paper to t. ruggeri on occasion of his .. th birthday ; he also thanks the organizers for a very stimulating and enjoyable meeting .
t.b . acknowledges hospitality at and support from the university of pavia during a working visit .
research supported in part by prin grant @xmath172 .
i. bakas , _ geometric flows and ( some of ) their physical applications _ , avh conference advances in physics and astrophysics of the 21st century , 6 - 11 september 2005 , varna , bulgaria , arxiv : hep - th/0511057 ( 2005 )
. t. buchert , _ on average properties of inhomogeneous fluids in general relativity : 2 .
perfect fluid cosmologies _ , gen .
* 33 * , 1381 ( 2001 ) .
t. buchert , _ dark energy from structure : a status report _
( special issue on dark energy ) , in press ; arxiv:0707.2153 ( 2007 ) .
t. buchert and m. carfora , _ regional averaging and scaling in relativistic cosmology _ , class .
* 19 * , 6109 - 6145 ( 2002 ) .
t. buchert and m. carfora , _ cosmological parameters are dressed _ , phys .
. lett . *
90 * , 31101 - 1 - 4 ( 2003 ) . m. carfora , a. marzuoli , _ model geometries in the space of riemannian structures and hamilton s flow _
quantum grav . 5 , 659 - 693 ( 1988 ) .
b. chow , s - c .
chu , d. glickenstein , c. guenther , j. isenberg , t. ivey , d. knopf , p. lu , l. ni , _ the ricci flow : techniques and applications : part i : geometric aspects _ , math
. surveys and monographs vol . *
135 * , am .
( 2007 ) .
ellis and t. buchert , _ the universe seen at different scales _ , phys .
. a ( einstein special issue ) * 347 * , 38 ( 2005 ) .
d. h. friedan , _ nonlinear models in @xmath173 dimensions _
physics * 163 * , no .
2 , 318419 ( 1985 ) . | ricci flow deformation of cosmological initial data sets in general relativity is a technique for generating families of initial data sets which potentially would allow to interpolate between distinct spacetimes .
this idea has been around since the appearance of the ricci flow on the scene , but it has been difficult to turn it into a sound mathematical procedure . in this expository talk we illustrate , how perelman s recent results in ricci flow theory can considerably improve on such a situation . from a physical point of view
this analysis can be related to the issue of finding a constant curvature template spacetime for the inhomogeneous universe , relevant to the interpretation of observational data and , hence , bears relevance to the dark energy and dark matter debates .
these techniques provide control on curvature fluctuations ( intrinsic backreaction terms ) in their relation to the averaged matter distribution . |
interest in the dynamical properties of open quantum systems at mesoscopic and atomic scales has lead to a rebirth of a form of scattering theory which was originally developed to deal with very complicated nuclear collision processes .
the origin of the scattering theory that we consider here came from the recognition that a collision between two nucleons , one of which is very heavy , can lead to the creation of an unstable , but very long - lived , compound nucleus which eventually decays . in the late 1930 s , kapur and peierls @xcite used this fact to formulate a nonperturbative approach to scattering theory in which the compound nucleus was viewed as a stable object which was made unstable by weak coupling to the continuum .
in the late 1940 s , wigner and eisenbud @xcite developed an alternate version of the kapur - peierls theory which lead to the very widely used r - matrix approach to scattering theory @xcite .
the idea behind both these theories is to decompose configuration space into an internal region ( reaction region ) and an external asymptotic scattering region .
as we shall see , this approach can be made clean and rigorous if there is an abrupt ( in configuration space ) transition between the scattering region and the asymptotic region .
the internal region can be modeled in terms of a complete set of states with fixed boundary conditions on the surface of the internal region .
the internal states can then be coupled to the external asymptotic states .
bloch @xcite and feshbach @xcite both showed that a consistent theory requires that the coupling between the internal and external regions must be singular . in the 1960 s , weidenmuller developed a phenomenological hamiltonian approach to nuclear scattering based on this picture @xcite .
the hamiltonian for the interior region was based on the shell model of the nucleus .
the hamiltonian for the exterior scattering region described the asymptotic states .
the strength of the coupling between the interior and exterior regions was an unknown input parameter .
the weidenmuller approach to scattering theory created a framework with which to study the manifestations of chaos in the scattering properties of nuclear systems , as well as mesoscopic and atomic systems .
it is now well known that chaos manifests itself in bounded quantum systems by affecting the statistical distribution of spacings between energy levels @xcite . for open quantum systems , in regimes where scattering resonances are not strongly overlapping
, underlying chaos affects the statistical distribution of resonance spacings and resonance widths @xcite .
if the hamiltonian of the interior region is chosen to be a random matrix hamiltonian , then predictions can be made as to the statistical properties of resonance spacings and resonance widths .
these predictions can then be compared to nuclear scattering resonances , or to resonances in electron waveguides [ @xcite ] , @xcite , and [ @xcite ] and electro - magnetic resonators [ @xcite ] and @xcite whose cavities have been constructed to yield classically chaotic behavior .
the agreement between random matrix theory predictions and numerical and laboratory experiments is quite good .
the weidenmuller hamiltonian is phenomenological and does not have information about the strength of the interface coupling .
the random matrix theories treat the interface coupling as an input parameter . in a real dynamical system , however , one needs a way to rigorously determine the coupling at the interface of the internal and external regions
. a very useful way to fix the coupling at the interface was provided by pavlov @xcite .
the idea of pavlov was that the surface coupling could be fixed by the requirement that the total hamiltonian ( including interior and exterior regions ) be hermitian . in this paper , we use a simple textbook scattering problem to illustrate and clarify many issues concerning this approach to scattering , and we then use it to study the effect of disorder on the scattering process . in section ( 2 )
, we describe the textbook " scattering problem and obtain exact expressions for the reaction function , the scattering function , and the wigner delay times . in section ( 3 )
, we develop the hamiltonian for interior and exterior regions of the scattering system . in section
( 4 ) we derive the hermiticity condition . in section ( 5 ) ,
we derive the reaction function . in section ( 6 )
, we derive the scattering function and locate resonance energies for the case of a smooth scattering potential . in section ( 7 ) , we use this theory to study the scattering of a particle wave from a random scattering potential . finally , in section ( 8) , we make some concluding remarks .
we will first consider the scattering of a particle of mass , @xmath0 , due to the potential shown in fig .
the quantum particle enters from the right with energy e and is reflected back to the left by an infinitely hard wall located at @xmath1 . a barrier of height , @xmath2 ,
is located @xmath3 .
the schroedinger equation , which describes propagation of a particle wave , @xmath4 , for all times , @xmath5 , is given by @xmath6 where @xmath7 is planck s constant and the potential , @xmath8 for @xmath9 , @xmath10 for @xmath11 , and @xmath12 for @xmath13 .
since the potential is infinite for @xmath14 , no wavefunction can exist in that region of space . the schroedinger equation for energy eigenstates @xmath15 , for region i ( @xmath3 ) in fig .
( 1 ) , is @xmath16 energy eigenstates in region i have the form @xmath17 where @xmath18 and @xmath19 is a normalization constant .
the schroedinger equation for energy eigenstates @xmath20 , for region ii ( @xmath21 ) in fig .
1 , is @xmath22 energy eigenstates in region ii have the form @xmath23 where @xmath24 .
the first term describes the incoming part of the energy eigenstate and the second term describes the outgoing part of that state .
the constant , @xmath25 , is a normalization constant and @xmath26 is the _ scattering function_. when @xmath27 , then @xmath28 and the incoming wave is reflected from the wall at @xmath1 with an overall phase shift of @xmath29 since we can rewrite the minus sign as @xmath30 .
thus , @xmath26 , contains information about the phase shift of the scattered wave due to the potential barrier , @xmath2 , in the region , @xmath3 .
we can determine how the two parts of the energy eigenstate are connected by the condition that the wavefunction and the slope of the wavefunction must be continuous at @xmath31 .
these two conditions can be combined into a statement that the `` logarithmic derivative '' of the wavefunction must be continuous at the interface .
thus we must have @xmath32 at the interface .
the function , @xmath33 , is called the _ reaction function _ and it will play an important role in the following sections .
the s - matrix is then given by @xmath34 where the reaction function , @xmath35 since the scattering function has unit magnitude we can also write it @xmath36 where @xmath37 is the phase shift of the scattered wave relative to the case for which @xmath27 .
it is straightforward to show that @xmath38 the time delay of the scattered particle due to its interaction with the barrier , @xmath2 , is given by @xmath39 this time delay is called the _ wigner delay time _ [ @xcite ] . for the scattering system we consider here , the phase angle ,
@xmath40 is given by @xmath41},\eqno(2.11)\ ] ] where @xmath42 and @xmath43 .
the wigner delay time is given by @xmath44 @xmath45 } { \biggr)}-{2ma^2\over { \hbar}{\beta}}.\eqno(2.12)\ ] ] in fig .
( 2.a ) we show the phase angle , @xmath37 , as a function of energy , and in fig .
( 2.b ) , we show the wigner delay time for parameter values , @xmath46 , @xmath47 , @xmath48 , and @xmath49 . the peaks in fig .
( 2.b ) occur at values of the energy where the particle wave resonates with the barrier region .
the fact that the wigner delay time goes negative for low energies means that when @xmath50 , the particle can be reflected earlier by the barrier , @xmath2 , than it would be if @xmath27 . in fig .
( 3 ) , we show that exact energy eigenstates as a function of position for a range of energies . at resonant energies ,
the energy eigenfunctions have enhanced probability in the scattering region .
these resonances are also often called _ quasi - bound states _ because they are associated with complex energy poles of the scattering function . in fig . ( 4 ) , we plot those complex energy poles of @xmath26 which give rise to the first three resonance peaks in the wigner delay time .
it is important to note that for scattering systems with higher space dimension , complex energy poles can have a profound effect on the scattering properties .
for example , in two dimensional electron wave guides they can completely block transmission in some channels [ @xcite ] , [ @xcite ] .
most of the interesting information regarding scattering events is contained in the energy dependence of the scattering phase shifts and in the location of quasi - bound state poles .
however , we generally can not find those quantities exactly as we did in the previous section . there are a number of techniques for finding them if the scattering potential is weak , but not if the scattering potential is strong . in subsequent sections ,
we describe an approach to scattering theory which can deal with strong interactions , provided they are confined to a localized region of space . as we mentioned in sect .
( 1 ) , this approach to scattering theory was originally developed to deal with nuclear scattering processes , but recently has provided an important tool for studying universal properties of scattering processes induced by underlying chaos .
we shall use this alternate approach to scattering theory to study the system considered in sect .
( 2 ) , and a similar system with a random scattering potential . in this way
, we can compare its predictions to the exact results which were obtained in sect .
let @xmath51 denote the position operator and let @xmath52 denote its eigenstates , so that @xmath53 .
the position eigenstates satisfy a completeness relation , @xmath54 , where @xmath55 is the unit operator , and they are delta normalized , @xmath56 . since the potential energy is infinite for @xmath57 , all states will be zero in that region .
we can divide the completeness relation into three parts , @xmath58 , @xmath59 , and @xmath60 , so that @xmath61 .
however , the operator @xmath62 acting on any state gives zero because the potential energy is infinite in that region .
therefore , we can remove @xmath62 from the completeness relation without changing our final results .
thus , from now on we will write the completeness relation as @xmath63 . the operators , @xmath64 and @xmath65 , are projection operators .
they have the property that @xmath66 , @xmath67 , and @xmath68 .
this is easily checked by explicit calculation .
the operator @xmath64 projects any state or operator onto the interval , @xmath69 , while the operator @xmath65 projects any state or operator onto the interval , @xmath70 .
in other words , if the state , @xmath71 has spatial dependence , @xmath72 , over the interval @xmath73 , then the state @xmath74 for @xmath75 and the state @xmath76 for @xmath13 . inside the region
i , @xmath75 in fig.(1 ) , we define a hamiltonian , @xmath77 where @xmath78 is the momentum operator and @xmath0 is the mass of the particle .
the hamiltonian , @xmath79 , is hermitian and therefore it will have a complete , orthonormal set of eigenstates which we denote as @xmath80 .
we can write the eigenvalue problem in the region , @xmath3 , as @xmath81 , where @xmath82 is the @xmath83 energy eigenvalue of @xmath79 and @xmath84 . because there is an infinitely hard wall at @xmath1 , the eigenstates @xmath85 must be zero at @xmath1 .
we have some freedom in choosing the boundary condition at @xmath31 and we will do that later .
the completeness of the states , @xmath80 , allows us to write the completeness relation , @xmath86 .
orthonormality requires that @xmath87 .
the hamiltonian for the region , @xmath88 , can be written @xmath89 its eigenvalues are continuous and have range , @xmath90 .
the eigenvector of @xmath91 , with eigenvalue , @xmath92 , will be denoted @xmath93 .
the eigenvalue equation then reads , @xmath94 .
any state , @xmath71 , can be decomposed into its contributions from the two disjoint regions of configuration space as @xmath95 .
we can expand @xmath96 in terms of the complete set of states , @xmath80 and we obtain , @xmath97 where @xmath98 .
the function @xmath99 must be equal to the function , @xmath100 , at the interface , @xmath31 .
in addition , the slopes of these two functions must be equal at the interface .
we couple the two regions of configuration space , at their interface , with the singular operator , @xmath101 .
the coupling constant , @xmath102 , will be determined later . then @xmath103 and @xmath104 it is useful to remember that @xmath105 , @xmath106 if @xmath107 , and @xmath108 if @xmath109 .
note also that @xmath110 and @xmath111 the total hamiltonian of the system can be written @xmath112 the energy eigenstates , @xmath113 , satisfy the eigenvalue equation @xmath114 .
the energy eigenstates can be decomposed into their contributions from the two regions of configuration space , so that @xmath115 where @xmath116 .
the eigenvalue equation then takes the form @xmath117 this yields a series of equations @xmath118 for @xmath119 and @xmath120 before we can proceed further , we must find conditions for hermiticity of the hamiltonian , @xmath121 .
consider the arbitrary states , @xmath71 and @xmath122 .
the condition for hermiticity of these states is that @xmath123 we can decompose the states @xmath71 and @xmath122 into their contributions to the two disjoint configuration space regions and write them in the form , @xmath124 where @xmath98 and @xmath125 where @xmath126 . in the interior region , we have expanded @xmath71 and @xmath122 in terms of the complete set of energy eigenstates , @xmath127 , of the hamiltonian , @xmath79 . to simplify the notation ,
let @xmath128 , @xmath129 , @xmath130 , and @xmath131 .
the states can be written @xmath132 then the condition for hermiticity is @xmath133\ ] ] @xmath134 + { \sum_j}[{\langle}{\xi}_p|{\hat h}_{pq}|{\psi}_j{\rangle } -{\langle}{\psi}_p|{\hat h}_{pq}|{\xi}_j{\rangle}^*]\ ] ] @xmath135=0.\eqno(4.6)\ ] ] we can now evaluate eq . (
4.6 ) term by term .
let @xmath136 , @xmath137 , @xmath138 and @xmath139 .
then @xmath140 @xmath141 @xmath142 is the hermiticity condition . for our subsequent discussion , it is useful to note that for the special case when the eigenstate , @xmath143 , has zero slope on the interface , so @xmath144 , the hermiticity condition reduces to @xmath145 @xmath146 if we let @xmath147 and @xmath148 in eq .
( 4.8 ) , the hermiticity condition reduces to @xmath149 where @xmath150 .
this hermiticity condition allows us to determine the coupling constant , @xmath102 . if we remember that @xmath151 , then continuity of the energy eigenfunction at @xmath31 requires that @xmath152 thus , the hermiticity condition allows us to determine the strength of the coupling at the interface .
we can use the formalism derived in sections ( 3 ) and ( 4 ) to derive an expression for the reaction function , @xmath33 . let us return to eq .
if we multiply by @xmath153 , use eq .
( 3.4 ) , and the normalization condition , @xmath154 , we obtain @xmath155 similarly , if we multiply eq .
( 3.12 ) by @xmath156 , we obtain @xmath157 @xmath158 where @xmath159 .
before we can proceed further , we must decide on boundary conditions for our states , @xmath143 .
we can use any complete basis set in the region , @xmath3 , provided they have the property @xmath160 .
we are free to choose the boundary condition at @xmath31 .
we have seen in eq .
( 4.8 ) that the hermiticity condition becomes especially simple for the boundary condition , @xmath161 . therefore , that is the boundary condition that we will use here .
the states with these boundary conditions are given by @xmath162 , for @xmath163 odd @xmath164 , and the corresponding eigenvalues are @xmath165 . the states , @xmath143 and the eigenvalues , @xmath82 , should not be confused with the exact energy eigenvalues and eigenstates in region i. the states , @xmath143 , only serve to provide us a complete set of states in region i which can then be used to construct the exact energy eigenstates in region i. with these boundary conditions , eq .
( 5.2 ) takes the form @xmath166 let us now show that this description of the system leads to well known results .
we can derive the reaction function originally obtained by wigner and eisenbud .
let us return to eq .
( 5.1 ) and solve for @xmath167 .
we find @xmath168 if we combine eqs .
( 4.9 ) and ( 5.4 ) , we obtain @xmath169 from eq .
( 5.5 ) , we can express the logarithmic derivative of the outside wavefunction , @xmath170 , evaluated at the interface ( and therefore the reaction function , @xmath33 ) , in terms of the inside states .
we find @xmath171 for scattering in higher dimensional space or for systems with internal degrees of freedom the reaction function becomes a matrix and is then called the _
reaction matrix_. in eq .
( 2.8 ) , we obtained an exact expression for the reaction function .
if we equate eqs .
( 2.8 ) and ( 5.6 ) , we obtain @xmath172 for @xmath163 odd .
let @xmath173 .
then we can write eq .
( 5.7 ) in the form @xmath174 from ref .
[ @xcite ] one can show that this is just the definition of the series expansion of the tangent function in terms of its argument .
the reaction function can also be expressed as a product of matrices .
let us consider the first @xmath175 eigenvalues and eigenvectors of the the hamiltonian , @xmath79 ( we later let @xmath176 ) .
we then write the hamiltonian , @xmath79 , in the matrix representation in which it is diagonal ( in this representation we denote it as @xmath177 ) , and obtain @xmath178 we also introduce the vector of eigenstates of @xmath79 , but evaluated at the interface , @xmath179 then the reaction function can be written in the compact form @xmath180 where @xmath181 is the @xmath182 dimensional unit matrix and @xmath183 denotes the hermitian adjoint of @xmath184 .
it is useful to express the scattering function , @xmath26 , in matrix notation , because we can then find a very interesting expression for the complex energy poles of @xmath26 . in eq .
( 2.7 ) , we obtained an expression for the scattering function in terms of the reaction function . in terms of the truncated reaction function
it is @xmath185},\eqno(6.1)\ ] ] where @xmath186 in table ( 1 ) , we give the positions of the three lowest energy quasi - bound state poles found by using the exact expression for @xmath33 given in eq .
these values are exact to the number of digits shown . also in table ( 1 ) , we give the positions of those poles found by using eq .
( 6.1 ) for four different truncations , @xmath187 , @xmath188 , @xmath189 , and @xmath190 .
the values obtained using eq .
( 6.1 ) converge slowly to the correct answer .
.comparison of pole positions with exact result [ cols="<,<,<,<,<",options="header " , ] the scattering function can be written in several other forms as well .
let us introduce the column vector , @xmath191 then the scattering function becomes @xmath192 } = { \rm e}^{-2ika } { \biggl[}1+{{2ik_n}\over { 1-ik_n}}{\biggr ] } .\eqno(6.4)\ ] ] where @xmath193 using eq .
( 6.4 ) , we can write ( we suppress the index @xmath175 ) @xmath194 @xmath195 @xmath196 the scattering function , @xmath197 , can now be written in the form @xmath198}{1\over
e{\bar 1}_n-{\bar h}^n_{in}}{\bar w}_n.\eqno(6.7)\ ] ] where @xmath199 is the adjoint of the matrix , @xmath200 , and @xmath201 denotes its determinant .
the poles of the scattering function are given by the condition @xmath202=0 .
\eqno(6.8)\ ] ] eq .
( 6.8 ) gives an @xmath203 order polynomial in @xmath92 whose solutions are the complex energies which locate the poles . at first sight , it would appear that the complex energy poles of the scattering function are simply given by the eigenvalues of the the non - hermitian matrix , @xmath204 .
however , the column matrices , @xmath205 depend on energy , @xmath92 , and therefore the eigenvalues , @xmath206 , of @xmath207 also depend on energy , @xmath92 . in fig .
( 5 ) , we locate the zeros of @xmath208= { \rm det}[e{\bar 1}-{\bar h}_{eff}-i{\bar w}{\bar w}^{\dagger}]$ ] in the neighborhood of the first three resonance energies .
the zeros satisfy the equation , @xmath209=(e-{\mu}_1(e))(e-{\mu}_2(e ) ) { \times} ... {\times}(e-{\mu}_n(e)).\eqno(6.9)\ ] ] in fig .
( 5.a ) we plot @xmath210 versus @xmath92 for complex values of @xmath92 in the neighborhood of the first pole .
we see that @xmath211 at the energy of the first pole . in figs .
( 5.b ) and ( 5.c ) , we plot @xmath210 and @xmath210 at the energies of the second and third poles , respectively .
again we see that they go to zero at the energies of their respective poles .
the potential used in section ( 2 ) , is a smooth step function which allows us to solve the scattering problem exactly
. it also allows us to obtain exact expressions for the basis state energies , @xmath212 and coupling constants , @xmath184 . for our simple scattering problem , matrix elements of @xmath184 simply alternate between two constant values .
we now use these same methods to study scattering from a random step potential .
the random potential we use is a sequence of 10 tent - like shapes ( upright or inverted ) on the interval @xmath3 with @xmath213 .
we again choose @xmath46 and @xmath47 .
we can express the potential in the form @xmath214 the values of @xmath215 are chosen at random from a uniform distribution over the interval , @xmath216 .
@xmath217 is the heaviside function .
one hundred different realizations of this potential are shown in fig .
6 . for this larger value of @xmath218 , there are about 10 to 12 scattering resonance peaks in the energy interval @xmath219 . to find the eigenvalues and eigenfunctions for this random potential we used an implicit finite difference method .
when there is no randomness ( @xmath220 for all @xmath163 ) , the wigner delay time can be calculated exactly as we showed in section 2 .
( 7 ) shows the wigner delay times when @xmath220 for all @xmath163 . in figure ( 7 ) , the exact result is given by the solid line , and the approximate result ( obtained by using the reaction matrix series expansion , eq .
( 5.6 ) ) is given by the discrete dots .
the agreement is excellent . in fig .
( 8) , we plot the values of @xmath221 which are obtained from the 100 hundred realizations of the random potential when @xmath213 .
the peak points ( positive and negative ) on the oscillating solid line give values of @xmath221 for the case @xmath220 for all @xmath163 .
we see that the higher eigenmodes are not affected by the random potential . in fig .
( 9 ) , we show the wigner delay times for 100 different realizations of random potential .
the positions of resonance peaks as well as their widths change with different potentials . in fig .
( 10 ) we plot a histogram of values of the wigner delay times shown in figure ( 9 ) .
this is similar to distribution one can get from rmt calculations , except that our distribution has a longer tail ( see @xcite @xcite ) . finally , in fig .
( 11 ) we show the distribution of wigner delay time resonance widths . we find that the distribution of delay time widths is fairly symmetrically distributed around its average value .
as we can see , this approach allows us to compute the scattering properties of a variety of shapes of scattering potential with great accuracy .
the only constraint is that the scattering potential must occupy a well defined region of space .
the configuration space scattering picture , described in sections ( 3)-(6 ) is a basis for weidenmuller s phenomenological hamiltonian theory of nuclear scattering , and it has provided a framework to look for universal behavior in the scattering properties of open quantum systems whose underlying classical counterparts are chaotic [ @xcite ] . there is now a large body of work [ @xcite ] which shows that bounded quantum systems , whose underlying classical dynamics is chaotic , have energy eigenvalues and energy eigenfunction which have statistical properties similar to those of certain types of random hamiltonian matrices .
similar questions are being asked about the scattering properties of open quantum systems for open systems , one can look at the statistical properties of the spacings and widths of scattering resonances , and of wigner delay times . the theory described in sections ( 3)-(6 )
can be used to compare the results of laboratory and numerical experiments to random hamiltonian matrix predictions .
for example , for scattering systems in which no magnetic fields are present , the hamiltonian , @xmath222 , which is formed from the eigenvalues of @xmath223 , is replaced by a hamiltonian , @xmath224 , which is formed from the eigenvalues of a hamiltonian matrix whose matrix elements are random numbers which are gaussian distributed .
the elements of the coupling matrix , @xmath225 , are also chosen from a random distribution .
the strength of the coupling constant , @xmath102 , is left as a variable parameter in these random matrix theories .
the random matrix theory predictions are in qualitative agreement with numerical experiments and some laboratory experiments . however , a number of assumptions underly the random matrix theory predictions , and there are a number of issues that remain open . for actual scattering systems ,
the coupling constant , @xmath102 , is fixed by the detailed dynamics using the hermiticity condition , but in random matrix theories it is a variable parameter . in random matrix theories , the coupling matrix , @xmath225 ,
is chosen from a random distribution .
how does it actually look for a deterministic system with underlying classical chaos ? this has never been studied .
random matrix theories also neglect the dependence of @xmath226 on the energy , @xmath92 .
the actual error made in making this assumption needs to be understood . for two dimensional electron waveguides
, there is an additional problem that at the threshold energy where a new channel opens , quasi - bound states inside the cavity can extend far down the leads .
@xcite it is not clear how well the theory presented in sections ( 3)-(6 ) and therefore the random matrix theory , can describe scattering in those regions .
never - the - less , this approach to scattering theory has allowed contact to be made between chaos theory and scattering theory for open quantum systems .
as we have just shown , it can also provide a powerful tool to study scattering processes in atomic and mesoscopic devices , when intrinsic disorder in the medium needs to be included .
the authors wish to thank the robert a. welch foundation grant no .
f-1051 , nsf grant no .
int-9602971 , and doe contract no .
de - fg03 - 94er14405 for partial support of this work .
we thank npaci and the university of texas at austin high performance computing center for use of their computer facilities .
we also thank german luna acosta and gengiz sen for helpful discussions .
huibers , s.r .
patel , c.m .
marcus , p.w .
brouwer , c.i .
duruoz , and j.s .
harris , jr .
, `` distributions of the conductance and its parametric derivatives in quantum dots '' phys . rev .
* 81 * 1917 - 20 ( 1998 ) .
fyodorov and h - j sommers , `` statistics of resonance poles , phase shifts and time delays in quantum chaotic scattering : random matrix approach for system with broken time - reversal invariance '' j.math.phys.bf 38 1918 - 1981 ( 1997 ) .
kyungsun na and l.e .
reichl , `` electron conductance and lifetimes in a ballistic electron waveguide '' , j. stat . phys .
* 92 * 519 - 542 ( 1998 ) ; kyungsun na and l.e .
reichl,effect of the quasibound states on quantum transport in a ballistic electron waveguide , phys . rev .
b * 59 * 13073 - 9 ( 1999 ) .
g. akguc and l.e.reichl,`evanescent modes in the reaction - matrix theory of electron waveguides ' , ( preprint , ilya prigogine center of statistical mechanics and complex systems , u. of texas at austin , june 2000 . ) | a version of scattering theory that was developed many years ago to treat nuclear scattering processes , has provided a powerful tool to study universality in scattering processes involving open quantum systems with underlying classically chaotic dynamics .
recently , it has been used to make random matrix theory predictions concerning the statistical properties of scattering resonances in mesoscopic electron waveguides and electromagnetic waveguides .
we provide a simple derivation of this scattering theory and we compare its predictions to those obtained from an exactly solvable scattering model ; and we use it to study the scattering of a particle wave from a random potential .
this method may prove useful in distinguishing the effects of chaos from the effects of disorder in real scattering processes . |
observations of distinct populations of stars in the milky way galaxy ( young , metal - rich stars in the disk and older , metal - poor stars in the halo ) led ( * ? ? ?
* els ) to propose that the galactic halo formed before the disk , during the rapid collapse of a gas cloud . in this picture , the stars in the halo are approximately the same age , with abundances which increase towards the center of the galaxy .
the model predicts a relatively smooth form and shape for the halo .
simple models , for instance , parameterize the distribution of matter in the galactic halo as a flattened spheroid , where the flattening is given by an axial ratio @xmath10 ( and @xmath11 if triaxiality is considered ) , with a power - law density profile @xmath12 .
as the positions and motions of luminous objects in the galactic halo in principle provide tracers of mass within the halo potential , study of these objects is important for understanding the distribution of dark matter as well as providing clues to the formation of the galaxy .
model parameters for this simple description of the halo have been obtained by empirical fits to the distributions of such astronomical objects as galactic globular clusters , k dwarfs , and a - colored stars . here
the term `` a - colored stars '' includes all types of stars , including blue horizontal branch ( bhb ) and blue straggler ( bs ) stars , which have colors and spectra similar to those of the bright , nearby , main sequence a stars on which the spectral classification was defined .
the distribution of galactic globular clusters yields a fit to the structure parameters of @xmath13 and @xmath14 @xcite .
@xcite used rr lyrae stars to find @xmath15 out to distances of 25 kpc , with a steeper decline thereafter .
@xcite also measured a distribution of rr lyrae stars with @xmath15 to at least 50 kpc .
hubble space telescope counts of k dwarfs by @xcite yield @xmath16 , and @xmath17 .
it is of course not required that @xmath18 remain constant with galactocentric radius .
@xcite counted bhb stars and found evidence that @xmath18 increases from 0.5 to 1 as one moves out to @xmath19 kpc with @xmath20 . using this changing @xmath18
, @xcite find @xmath21 for rr lyraes .
@xcite find a flat @xmath18 , but less evidence for changing flattening with galactocentric radius . as part of an extensive kinematic analysis of several hundred nearby halo stars with [ fe / h ] @xmath22 , @xcite find that the axial ratio increases from 0.65 at @xmath23 kpc to 1.0 at @xmath24 kpc .
most of these studies use a few hundred objects at most on which to base their results , and only a few , such as @xcite , have kinematic information .
in contrast to the els model , recent research suggests that some or all of the halo may have been formed through accretion of smaller dwarf galaxies ( see * ? ? ? * ; * ? ? ? * and references therein ) .
one example of a dwarf galaxy thought to be in the process of accreting has been discussed by @xcite .
as these smaller galaxies fall into the milky way , they may be tidally disrupted and break up into long streams of stars which can persist for 1 - 10 gyr @xcite .
a - colored stars at faint magnitudes are usually assumed to be halo population ( pop ii ) stars , and hence bhb stars .
however , it has become clear that many distant blue stellar objects have hydrogen line widths indicative of main sequence gravity a stars rather than the less massive bhb stars @xcite .
why ca nt all of these high - gravity a - colored stars be normal main - sequence stars ? the well - known arguments against this are two - fold : first , except for dwarf irregular companions to the milky way such as the lmc and the sextans dwarf , there is no evidence that significant star formation has taken place in the far halo ( r @xmath25 30 kpc ) within the last billion years , which is greater than the main - sequence lifetime of an a star .
second , the stars are too far from the star forming regions in the galactic plane to have reached their present positions in the far halo with any reasonable assumed velocity ( it takes 300 myr to travel 30 kpc at 100 km s@xmath26 ) .
blue straggler stars in globular clusters , which have ages closer to 10 gyr , are an interesting example of high surface gravity a - colored objects which are present in an old environment .
these high surface gravity bs stars are generally thought to be main sequence stars which result from mergers of , or mass transfer between , two lower mass stars @xcite .
whatever their formation mechanism , they should also be present at some level in the field of the galactic halo @xcite .
the sloan digital sky survey ( sdss ) is a large , international collaboration set up to survey 10,000 square degrees of sky in five optical passbands and to obtain spectra of a million galaxies , one hundred thousand quasars , and tens of thousands of galactic stars .
the data are being taken with a dedicated 2.5 meter telescope located at apache point observatory ( apo ) near cloudcroft , new mexico .
the telescope has two instruments : a ccd camera with 30 2k x 2k ccds in the focal plane , and two 320-fiber double spectrographs .
the imaging data are tied to a network of brighter astrometric standards ( which would be saturated in the main imaging data ) through a set of twenty - two smaller ccds in the focal plane of the imaging camera .
a 0.5 m telescope at apo will be used to tie the imaging data to brighter photometric standards . the usno 1 m telescope at flagstaff , az , has been used to define the photometric system , defining sdss standard stars in @xmath27 and @xmath28 ( smith et al . , in preparation ) . since the precise calibration for the sdss filter system is still in progress ,
magnitudes in this paper are quoted in the @xmath29 system , which approximates the final sdss system .
these systems differ absolutely ( with negligible color terms ) by only a few percent in @xmath30 , and no more than 10% in @xmath31 .
see @xcite , @xcite and @xcite for further information on the sdss , the filter system and the imaging camera .
the sdss has imaged a large fraction of the celestial equator ( @xmath32 ) as part of telescope and imaging camera commissioning .
catalogs produced from nearly 400 square degrees of these data contain thousands of a - colored stars suitable for probing the halo , with absolute photometry good to @xmath33% for objects with @xmath34 . for reference , blue stars with @xmath35 ( including a stars ) , have a sdss @xmath36 magnitude approximately equal to their johnson @xmath37 magnitude .
the sdss camera contains six columns of ccds , each of which records data for a 0.23 degree wide strip of the sky .
the ccds are continuously read out as the camera scans a great circle across the sky at the sidereal rate , so the length of an imaged strip grows at 15 degrees per hour .
the effective exposure time in each of the 5 filters is 54 seconds .
a `` run '' generates six parallel continuous strips of sky with a gap of 0.2 degrees between columns .
a second `` run , '' offset slightly from the first , is obtained to fill in the sky between the first six scan lines and fully image a 2.5 degree wide `` stripe '' on the sky .
figure [ fg1 ] shows @xmath38 vs. @xmath39 for all stellar objects with @xmath40 in 13.3 square degrees of one column of sdss data .
objects in different magnitude ranges are plotted in different colors . as discussed by @xcite , the position of the blue end of the galactic stellar locus shifts as a function of magnitude due to the sampling of populations of stars with different characteristic metalicities at different scale heights in the disk .
the temperatures as a function of color can be estimated by comparison with @xcite .
the stars in the a box in the diagram are expected to contain nearly exclusively hot stars with a strong balmer jump bhb stars , bss , and main sequence stars with spectral type a. these bluer objects do not shift by large amounts in color with magnitude .
boxes for quasar candidates , f , and g type stars are also indicated and will be discussed below .
all of the magnitudes used in this paper are point spread function ( psf ) fitted magnitudes for stellar objects , generated by summing the psf - weighted flux for each object ( lupton et al .
, in preparation ) .
saturation sets in for a - colored objects brighter than @xmath41 .
all magnitudes in this paper use the modified flux designation for faint magnitudes described in @xcite , though for the range of magnitudes relevant here , @xmath42 , the distinction is negligible . the boundaries of the a - colored star box were chosen to provide a sample of a stars , bhbs and bss , free from contamination from other populations of objects even to magnitudes as faint as @xmath43 .
the choice of a @xmath44 red limit was conservatively chosen in order to ensure that the gaussian tail of color measurement errors of the population of the very large numbers of f stars did not begin to scatter into the a star box at faint magnitudes .
all magnitudes and colors in the present work have been corrected for reddening using the maps of @xcite .
since all of the a - colored objects sampled here are at distances @xmath45 from the plane of the galactic disk , where most reddening - producing dust occurs , the full reddening correction of @xcite is applied to all objects . for this analysis , two samples of a - colored stars are chosen from two stripes of data .
the northern sample ( north of the galactic plane on the celestial equator ) is a filled stripe from sdss runs 752 and 756 , obtained 1999 march 21 and 22 .
the stripe is 87 degrees long ( @xmath46 [ j2000 ] , and 2.52 degrees wide ( @xmath47 ) , with 219 square degrees of data .
data in the region @xmath48 from run 752 was excluded , since the seeing increased sharply to @xmath49 in this region ( compared to typical seeing of @xmath50 for the rest of the scan ) . in galactic coordinates , the northern sample runs in a long arc from [ l , b ] = [ 236,37 ] up to [ 303 , 63 ] and then gradually back to [ 4 , 43 ] .
the northern sample contains 3126 point sources ( extended sources are rejected ) in the runs 752 and 756 which have dereddened colors in the box in figure [ fg1 ] ( @xmath51 and @xmath52 ) , and which have @xmath31 errors smaller than 30% ( corresponding to a limit of @xmath53 ) . the southern sample ( south of the galactic plane on the celestial equator ) is a filled stripe from sdss runs 94 and 125 , obtained 1998 september 19 and 25 .
the stripe is 60 degrees long ( @xmath54 ) and 2.5 degrees wide ( @xmath55 ) , with 151 square degrees of data .
the southern sample contains 1082 stars , to @xmath56 , with the same selection criteria as the northern sample .
the southern sample runs from [ l , b ] = [ 85,-57 ] through [ 125,-63 ] and then back to [ 184,-43 ] .
the sdss has imaged the majority of these equatorial regions twice .
sdss runs 259 , 273 and 745 ( obtained 1998 november 17 , 19 and 1999 march 20 , respectively ) , nearly completely overlap runs 94 , 125 and 756 , and are used to verify the findings in the primary runs .
the list of a - colored stars from runs 745 and 756 were matched ( without the requirement that the error in @xmath31 be smaller than @xmath57 ) , based on position ( within one arcsecond , well in excess of typical one sigma astrometric errors of 100 mas ) .
additionally , objects from these overlapping scans were matched without regard to color in order to estimate completeness , detection limits and reproducibility of the data .
the upper curve in figure [ fg2 ] shows the fraction of objects matched as a function of @xmath36 for all stellar sources .
the lower and noisier curve is that for matched objects within the a - colored box only .
the maximum level of @xmath58% matched in the lower curve is primarily due to objects in one sample or the other scattering out of the box .
one can estimate this fraction by noting that most of the objects in the box are on the red side , giving the box an effective width of 0.15 mag in @xmath39 .
when combined with typical color errors of 2% for two measurements , one derives an expected fraction @xmath59 .
the matched fraction decreases rapidly for @xmath60 . here
, the requirement that the object fall within the box in @xmath38 becomes the determining factor . from figure [ fg2 ]
, one sees that the a - colored star sample is uniform ( in the sense that it samples the same population of stars ) to @xmath61 and has detections to about @xmath53 ( 50% limit ) in the northern stripe .
after @xmath61 the population of stars includes an increasing fraction of stars slightly redward of the box whose photometric errors have scattered them into the box .
the fact that the upper curve does nt start to drop until @xmath53 indicates that the uniform sample is not affected by the detection limit , and is in that sense complete . a related procedure in the south yields a uniform sample to @xmath62 and detections to @xmath63 .
horizontal branch stars , especially those in a restricted color range , are approximately standard candles . a bhb absolute magnitude of @xmath64 can be used to infer approximate heliocentric distances from magnitudes for these objects .
this absolute magnitude was calculated by converting the v absolute magnitude of halo rr lyrae stars from @xcite into the @xmath36 filter , using the transformations from @xcite for the colors of rr lyrae stars .
since the @xmath36 magnitudes of rr lyrae stars in pal 5 ( figure [ fg7 ] ) are similar to the bhb magnitudes in pal 5 , the assumption has been made throughout this paper that the absolute magnitudes of bhbs are the same .
this approximation is dependent on the blue response of the filter used .
note , for example , that @xcite , find that for objects with colors at the blue end of the a star selection box , bhbs are about 0.9 magnitudes fainter in v than the rr lyrae magnitude ( this corresponds to a difference of 0.7 magnitudes in @xmath36 , given the transformations in * ? ? ? * ) .
figure [ fg3 ] presents a wedge diagram with right ascension and @xmath36 magnitude for all 4208 a - colored stars from both north and south samples .
significant structures are immediately apparent . the clump at [ l , b ] = [ 242,43 ] , or @xmath65 , and @xmath66 is the northern half of the sextans dwarf irregular galaxy @xcite at @xmath67 kpc ( the southern half of sextans is at @xmath68 and is not covered in this data set ) .
one also notes two dramatic arcs of structure at @xmath69 and @xmath43 in the north ( @xmath70 ) , and similar , though less pronounced , structure at @xmath71 and @xmath61 in the south ( @xmath72 ) .
the large structure in the north was detected by @xcite in a search for variable rr lyrae stars in the overlapping sdss runs 745 and 756 taken two days apart .
it contains a remarkable population of @xmath73 rr lyraes clustered in the same way as the bluer a - colored stars in the northern sample .
there are numerous smaller apparent concentrations throughout this diagram .
the galactic globular cluster , pal 5 , for example , has its horizontal branch at @xmath74 , @xmath75 in the north .
an enlargement of the three degrees of arc in the pal 5 area of sky is shown to the right in figure [ fg3 ] .
along with the stars of pal 5 s bhb at @xmath74 , the narrow sequence of pal 5 bss ( see [ pal5example ] ) is clearly visible from @xmath76 , also at @xmath75 .
a number of straight - forward checks are performed to verify that the large structures in figure [ fg3 ] are not artifacts . in order to assess whether these structures could be due to selection effects ,
low - redshift quasar candidates are color selected from the sdss survey from a color box with boundaries @xmath77 , @xmath78 ( the small q box in figure [ fg1 ] ) and plotted in figure [ fg4 ] .
figure [ fg4 ] shows no significant magnitude ( and thus implied spatial ) structure to @xmath56 , except for the increasing density of quasar candidates to fainter magnitudes . while not every object in this color box is a quasar , other hot blue sources which contaminate the low redshift quasar population , such as white dwarfs and narrow emission line compact galaxies , also have population distributions which are expected to be isotropic , showing no structural signature of the galactic disk , bulge or halo @xcite .
the uniformity of the distribution in figure [ fg4 ] demonstrates the uniformity of the photometry and calibration in the sdss samples to magnitude limits of @xmath56 for objects with colors similar to a stars .
fainter than @xmath79 , the errors in the @xmath31 magnitudes of f stars become large enough that some of them scatter into the quasar color box .
the distribution of f stars in the northern and southern sample areas is discussed in [ fgstars ] .
the photometric calibration within an individual stripe is accurate to a few percent due to the continuous drift scanning - method of data collection and photometric observing conditions .
relative photometric accuracy of objects detected in multiple scans is good to @xmath80 for blue objects to @xmath61 in the @xmath81 and @xmath82 filters . the absolute calibration , and the calibration between runs observed in different seasons , relies upon calibration patches obtained with a small photometric telescope , and is not yet final . to verify that the absolute calibration is essentially identical in the northern and southern stripes ,
the density of quasar candidates ( chosen from the larger q box in figure [ fg1 ] to increase statistics ) as a function of magnitude for the north and south is plotted in figure [ fg5 ] .
the excellent agreement between the two data sets to @xmath83 is further evidence that the photometric calibration is uniform across the sky .
one notes from the earlier falloff for southern quasar candidates in figure [ fg5 ] that the southern sample completeness limit is about 0.5 magnitudes brighter than that of the northern sample . as the objects are selected in the bluest sdss filters , and since the @xmath84 filter is the least sensitive filter , completeness limits and the magnitude limit for detections of objects in the present work
is determined primarily by where the error in the @xmath31 band flux for objects becomes large .
note that although @xmath39 colors for quasars and a stars are similar , the @xmath38 colors of quasars are bluer than a stars by about one magnitude .
therefore , one expects to see a drop in a - star counts one magnitude brighter than the drop in quasar counts , as is verified by comparing the upper curve in figure [ fg2 ] to the northern points in figure [ fg5 ] .
the duplicate equatorial runs 745 and 273 were used to generate a plot similar to that shown in figure [ fg3 ] .
the structures noted in [ distrib ] are also clearly visible in figure [ fg6 ] , though there are fewer stars since only half of a filled stripe is plotted .
note that run 745 does nt cover sextans in ra , but does extend beyond the end of the northern structure , and thus more clearly defines its full length than in figure [ fg3 ] .
while the sdss continues commissioning of its spectrographs , which will obtain thousands of spectra of type a and bhb stars in addition to its main galaxy and quasar redshift survey , a small number of representative spectra have been obtained with other telescopes to investigate the nature of the selected a - colored objects .
the observations produced spectra for a sample of 6 a - colored stars selected from the sdss samples , along with spectra of 5 a - colored reference stars with known surface gravities .
the lick observatory 3 m telescope was used to obtain reference stars , including cs 29516 - 0011 , a bhb identified as a field horizontal branch star by @xcite ; wd 0148 + 467 , a da white dwarf identified by @xcite ; and hd 13433 , a known a0v star classified by @xcite , all with @xmath85 .
these spectra were obtained with exposures between 5 s and 60 s on the night of 1998 dec 19 at a resolution of about 6 , obtaining good s / n ratios of about 50:1 in the continua .
figure [ fg7 ] shows the color - magnitude diagram of pal 5 , generated from reddening - corrected sdss data .
stars of all colors within @xmath86 arcminutes of the center of the globular cluster pal 5 were selected ; note that there is significant field contamination here at [ l , b]=[1,46 ] .
the vertical lines show the @xmath39 limits of the a - colored star box . a 5400 s exposure of one blue straggler and a 1800 s spectrum of a bhb star ( marked with circled squares in fig . [ fg7 ] )
were taken at 3 resolution with the dual imaging spectrograph ( dis ) on the arc 3.5 meter telescope at apache point observatory on the night of 1999 may 11 .
proper motion studies confirm that both objects are members of the pal 5 globular cluster ( k. cudworth , private communication ) .
the resolution of these spectra is approximately 3 at a s / n @xmath87 for the brighter object and 15 for the fainter .
wavelength calibrations were obtained and standard flats and reductions were applied to all spectral data .
six additional 30 min spectra of @xmath88 a - colored stars selected from the sdss were observed on the nights of 1999 may 8 , 10 , and 11 with the dis at 3 resolution and s / n of @xmath89 .
the six spectra all confirmed that the a - colored star sample indeed consists entirely of blue stars with strong balmer lines , and is not contaminated by significant numbers of white dwarfs , quasars or other extragalactic objects such as compact e+a galaxies .
the primary diagnostic used to distinguish bhbs from a main sequence ( or blue straggler ) stars is the widths of the balmer line profiles as indicators of stellar surface gravity ( see * ? ? ? * and references therein ) .
main sequence a stars ( @xmath90 ) have @xmath91 , while horizontal branch stars , which have much smaller mass ( @xmath92 ) and larger radii @xmath93 , have @xmath94 .
this difference in surface gravities clearly shows up in the stark pressure - broadened balmer line profiles .
white dwarfs , which have much higher surface gravities typically @xmath95 , are easily distinguished from main sequence a stars . since white dwarf
@xmath38 colors are significantly bluer than those of a and bhb stars , they are not present in the a - star selection box in significant numbers .
none of the spectra of a - colored stars presented here is consistent with a white dwarf identification .
the standard indicator which separates main sequence gravity a stars from lower surface gravity bhb stars is the width of the h@xmath96 line at @xmath97 of the continuum .
one typically uses the h@xmath96 line because the a star flux is higher here than in longer wavelength balmer lines , and it is easier to fit the continuum at h@xmath96 than at shorter wavelength balmer lines .
the width at @xmath97 of the continuum is a measurement robust to the presence of metal lines which can apparently widen the absorption lines near the continuum , and is insensitive to a range of instrumental resolutions , which can change the measured depth of the line .
iraf was used to extract the 1d spectra , select a region of the spectrum containing the h@xmath96 and h@xmath98 lines , divide the continuum by a best fit 6@xmath99 order legendre polynomial , and then boxcar smooth and wavelength shift the spectra to match the resolution and redshift of the lick reference spectra .
the difference in resolution between the sets of observations is not critical , since the broad balmer lines typically have equivalent widths of 30 .
it must be noted that with very wide balmer lines it is difficult to uniformly fit the continuum in such a way that the lines themselves do not affect the fit .
the results are plotted in figures [ fg8]-[fg9 ] .
figure [ fg8 ] shows the pal 5 stars compared to the blue reference templates . the bhb star in pal 5 is a good match to the template bhb star .
the spectrum of the pal 5 blue straggler has much larger line widths consistent with the template a0v star .
the six additional representative a - colored objects were spectroscopically confirmed with spectra shown in figure [ fg9 ] .
for these six brighter ( @xmath100 ) sdss northern sample targets observed from apo in figure [ fg9 ] , 4 have narrow widths , one has a broad width ( lower left panel ) , and
one is apparently a metallic - lined am star ( upper left panel ) .
table 1 summarizes the spectroscopic targets , listing sdss i d ( or other i d such as hd number ) , @xmath101 , @xmath96 , dereddened @xmath36 magnitude , @xmath38 , @xmath39 , @xmath102 , l , b , and classification determined from the balmer line widths in the spectrum .
the radial velocities of the stars were only approximately measured ( with typical error of 30 @xmath103 ) , and ranged from -120 to @xmath104 , consistent with halo objects .
the resolution of the observations and uncertainties in the wavelength calibrations made it difficult to draw conclusions about clustering of this small sample of objects from radial velocities .
kinematic information will of course be an important future diagnostic of sdss a - colored stars in continuing searches for coherent halo structures .
as previous authors @xcite have noted , and the spectroscopy in this paper also shows , one must be careful drawing conclusions about galactic structure from color - selected a - stars in the halo of the milky way , since the blue star population is not homogeneous ; it includes at least two types of objects .
one gains insight into the nature of the a - star sample by examining the appearance of the globular cluster pal 5 in the star distribution in figure [ fg3 ] at @xmath105 .
pal 5 , located about 20 kpc from the sun , contributes 5 bhb stars at @xmath106 , and at least 7 bs stars at @xmath107 ( see figure [ fg7 ] ) . on an absolute scale ,
the bhbs in pal 5 have @xmath108 , and the bss are thus at absolute magnitudes of @xmath109 .
the large parallel arcs of a - colored stars present in figure [ fg3 ] are also separated by about 2 magnitudes , suggesting that they might represent associated bhb and blue straggler stars at a common distance .
if this assumption is correct , then one can use the location of the bhb in @xmath36 and the absolute magnitude of bhb stars to determine a distance to the stars in the arcs .
for example , the northern arc , with an implied bhb magnitude of @xmath110 , is then approximately 45 kpc from the sun . in order to support the identification of the stars in the overdensities
, one would like to be able to separate all of the high surface gravity stars from the low surface gravity stars .
presently lacking individual spectra for such a large sample , one turns to photometric multi - color indicators of surface gravity in the sdss filters . from a sample of 1121 a - colored stars , @xcite
show that they can separate ( although not perfectly ) high surface gravity stars from low surface gravity stars in @xmath111 color - color diagrams . due to the lack of a single - valued transformation between the @xmath112 and @xmath31 ( they have different behavior around the balmer jump ) , a simple conversion between @xmath38 and @xmath113 is not well defined .
thus , although one can not precisely convert the @xmath111 color separation criteria to the sdss filter system , a similar idea is applied to approximately separate the populations : extreme balmer jump bhb stars are in general slightly bluer in @xmath39 and redder in @xmath38 than those with higher gravities .
the other sdss filters available ( @xmath114 ) may be considered for color separation of gravities .
for instance , @xcite show that in synthetic a stars derived from kurucz model spectra @xcite , different gravities separate well in the @xmath28 filter , and they derive a direction in sdss color space that results in maximum separation . since @xmath82 magnitudes accurate to 2% are lacking , for most of the faint a star sample considered here , the stars were separated in @xmath38 and @xmath39 only .
( the @xmath115 color contributes little to surface gravity separation for objects with a star effective temperatures . )
figure [ fg10 ] shows the colors of the @xcite models , along with the colors of the six stars for which spectra exist , and the colors of the bss and horizontal branch stars in pal 5 .
the positions of the pal 5 stars plotted in figure [ fg10 ] are circled in the pal 5 color - magnitude diagram in figure [ fg7 ] .
one star which appears to be on the horizontal branch of pal 5 and also in the a - star box is far enough from the center of the globular cluster that its membership is questionable , so it was not included in figure [ fg10 ] .
the adopted separation between bhbs and high gravity stars ( thick line in figure [ fg10 ] ) in @xmath116 colors was based empirically on the spectra and photometry obtained here , guided by @xcite and @xcite .
the models are valid for stars with @xmath117 .
a few candidate rr lyraes , selected as in @xcite , are also indicated . applying the color separation in figure [ fg10 ] ,
the 3126 stars in the northern sample are split into 2041 candidate blue straggler stars and 1085 candidate bhb stars .
the southern sample split into 674 candidate blue straggler stars and 408 candidate blue horizontal branch stars .
the wedge diagrams for these separated stars are shown in figure [ fg11 ] ( bhbs only ) and figure [ fg12 ] ( bss only ) .
these plots separate and enhance the contrast of the structures found in figure [ fg3 ] , supporting the identifications in 4.1 .
the structures are interpreted as an elongated arc of bhb stars at @xmath118 in the north , and an arc of bs stars about 2 magnitudes fainter in the same direction in the north .
additionally , an arc of bs at about @xmath61 is apparent in the south .
both the northern and southern structures extend over 30 degrees or more of the sky .
the sdss stripe is @xmath119 wide in declination and so these large structures may extend well beyond the narrow stripes observed here .
one also detects the candidate bhb stars associated with the candidate bs stars in the southern structure .
figure [ fg13 ] shows histograms of the candidate bs and bhb stars in the ra ranges of the northern and southern structures as a function of magnitude .
while there are no significant structures apparent at @xmath120 in the north ( left panels ) , the plot clearly shows an overdensity of bhb stars at @xmath110 in the north at @xmath121 ( middle panels ) and also shows the overdensity in bhbs at @xmath122 in the south ( right panels ) .
the bs / bhb color - color separation is far from razor sharp .
figure [ fg11 ] , for example , which should show the absence of the second fainter arc from the candidate bhb star set and leave only the brighter arc in the north , does not do this completely .
this is expected , since the color separation technique does not definitively separate the two populations of stars ( see figure [ fg10 ] ) .
the separation is less effective at fainter magnitudes , where the errors are larger , and at redder colors , where the populations overlap in color - color diagrams .
the @xmath39 color distribution of candidate bs and bhb stars in the large structures in the north and south is histogrammed in figure [ fg14 ] . in the upper panel ,
candidate bss with @xmath123 and @xmath124 in the northern structure are indicated with a light line against @xmath39 color , while candidate bhbs ( @xmath125 ) in the same area are histogrammed with a heavy line .
the lower panel shows the southern structure objects with @xmath126 bss and @xmath127 bhbs . in this color system ,
an a0v main sequence star has @xmath128 , whereas a somewhat cooler a5v star will have @xmath129 .
the distribution of bs stars in figure [ fg14 ] shows that the bs candidate stars are consistent with intermediate ( a3-a5 ) a star colors , and the population is not dominated by the hottest a0 colored objects .
the bs candidate sample is also notably consistent with the pal 5 bs objects of figure [ fg7 ] , which also does nt contain any bs stars as hot as an a0 dwarf .
one might naively expect the a colored stars in the sextans dwarf irregular galaxy to represent its horizontal branch .
however , @xcite show that this galaxy has a young very blue main sequence ; if there is a horizontal branch , it contains very many fewer stars .
the ratio of the number counts of bs : bhb candidates is 36:19 for the sextans dwarf .
this ratio is similar to the ratio of the excess of bs candidates at @xmath61 to the excess of bhb candidates in the same magnitude range in the southern structure . in both cases ,
we assume that the candidate bhbs are misidentified high surface gravity stars . thus for sextans , nearly all the bs and bhb candidates are simply young main sequence stars .
the tight concentration of bhb magnitudes in figure [ fg13 ] and the existence of associated bs candidate stars are strong evidence that the northern and southern structures are real halo substructure .
however , one must explore the possibility that these structures could be artifacts of a smoothly varying galactic structure combined with the magnitude limits of the sdss samples .
one is concerned with the possibility that the northern structure might be associated with the bulge of the milky way , since it occurs at a galactic longitude of zero degrees .
figure [ fg15 ] shows a simulated halo density distribution expected for a smooth spheroidal distribution of bhb stars , assuming @xmath130 , @xmath131 , and @xmath132 .
one can see from this figure that the highest density of bhbs is closer to the galactic plane in the equatorial stripe ( indicated by the line of nodes in figure [ fg15 ] ) , than to galactic longitude of zero .
this is because the plane of the celestial equator passes closer to the center of the galaxy at the galactic plane than at galactic longitude of zero .
the halo density distribution depends only on the distance from the galactic center for a spherically symmetric halo population .
one must resort to an extremely unlikely model to reproduce an over - density such as that seen in figure [ fg3 ] within a smooth spheroid .
for example , assuming an unrealistic absolute magnitude of @xmath133 for the a - colored stars , and an extremely prolate ( @xmath134 ) bulge , one can obtain an overdensity near @xmath135 , but it will be much more spread out in magnitude than the tight bhb locus seen at @xmath110 .
additionally , such an exercise can in no way reproduce the observed overdensity in the south with the same smooth model .
the southern structure is nearly directly opposite the galactic center .
figure [ fg16 ] shows a wedge plot , similar to figure [ fg3 ] , but for f star candidates ( objects with @xmath136 ) , which are intrinsically about 3 magnitudes fainter than bhb stars .
one sees here the familiar overdensity in the southern sample , but the typical magnitude , @xmath43 , is about one magnitude fainter than that of the presumed bss of figure [ fg12 ] .
this indicates that one may be detecting f stars associated with the southern structure ( @xmath137 ) .
note that although the density of the f star concentration appears to fall off near @xmath138 , the density of qsos in figure [ fg4 ] also drops off in this region at about @xmath139 magnitude in @xmath36 , indicating that the sample is less complete in that region .
the rest of the f star distribution is consistent with a smooth galactic model .
( imagine shifting the model distribution in figure [ fg15 ] three magnitudes fainter . )
the halo f stars associated with the northern structure are too faint to be unambiguously detected with the sdss .
a magnitude - ra wedge plot for color - selected g stars for the same sdss samples is presented in figure [ fg17 ] .
it is reassuring to see no signs of the northern or the southern structures in this wedge plot ; the g stars ( @xmath140 ) associated with these structures are too faint to appear above the sdss magnitude limits .
this population may prove valuable for studies of galactic structure at closer distances than those probed by the bhb / bs stars .
a slight leak of f stars in to the g star box may be manifesting itself as structure in the south at @xmath83 .
as horizontal branch stars have @xmath108 and bs stars have @xmath141 , one can estimate the heliocentric distances to the largest structures using the simple magnitude - distance relation , @xmath142 .
assuming the sun is @xmath143 kpc from the galactic center , one can then use [ l , b ] for each object and calculate a galactocentric distance for each star .
the resulting galactocentric distance estimates are plotted as ( @xmath101 , @xmath144 ) polar wedge diagrams in figure [ fg18 ] ( for the bhb subsample ) and figure [ fg19 ] ( for the bs subsample ) .
the implied distance of the large clump of stars in the north is 46 kpc from the galactic center , with galactic rectangular coordinates extending from ( x , y , z ) = ( 12 , -13 , 41 ) to ( 26 , 1 , 34 ) centered at ( x , y , z ) = ( 21 , -5 , 38 ) [ units are kpc , where the sun is at ( -8,0,0 ) ] . the most significant clump in the south is inferred to be about 33 kpc from the gc extending from ( x , y , z ) = ( -18 , 9 , -25 ) to ( x , y , z ) = ( -25 , 3 , -22 ) centered at ( -22 , 6 , -24 ) . the significance of other smaller clumps seen in figures [ fg18 ] and [ fg19 ] is tantalizing , and merits kinematic followup to determine if they are moving groups or streamers . in this coordinate system
, sextans is at ( -36 , -53 , 58 ) and pal 5 is at ( 7 , 0 , 16 ) . because of the small declination coverage currently available , the full extent of these large halo structures is not known .
however , an extent of 30 degrees in ra at 46 kpc , with a ` depth ' of at up to ten kpc ( from the spread in magnitudes in figure [ fg13 ] ) , suggests a structure of extent at least 20 kpc by at least 2 kpc by up to 10 kpc . a significant fraction of the assumed width from the magnitude spread could be due to the intrinsic spread in the absolute magnitudes of bhb stars , and the actual structures could in fact be extremely long and narrow `` streams '' .
one can estimate the implied total mass for the large structures in the northern and southern samples from the number of bhb stars observed . from the data associated with figure [ fg13 ] ,
one estimates that there are 200 bhb stars ( above background ) in the northern structure .
to arrive at this estimate , one uses the counts per square degree of bs and bhb stars in the left hand panel in figure [ fg13 ] to estimate the background ; the counts above background are assumed to be associated with the structures .
excess counts at @xmath110 in the northern sample and at @xmath145 in the southern sample were assumed to be bhb stars .
likewise , excess counts at @xmath43 in the northern sample and at @xmath61 in the southern sample were assumed to be bss . in pal 5 , there are five bhb stars in the sample . from comparisons of the total magnitude of pal 5 to other globular clusters
as tabulated in @xcite , we estimate the mass of pal 5 to be @xmath146 .
scaling up from a globular cluster mass , a lower limit on the inferred mass in the part of the northern structure detected is then @xmath147 .
this is roughly consistent with estimates of the mass from extrapolations from the number of rr lyrae stars .
ivezic et al .
( 2000 ) detect nearly 80 rr lyrae variables at @xmath148% efficiency within the northern clump , as compared to five in pal 5 , giving a ratio of 160 to five as compared with the above ratio of 200 to five .
an estimated 60 bhb stars in the southern structure implies a lower limit for the mass of @xmath149 .
there is of course no way to tell what fraction of the structure is detected ; more data from the sdss survey will eventually be able to map out the full extent of such structures .
it is interesting to calculate the ratio of bs : bhb stars in the large structures . in the globular cluster pal 5 , in the color range @xmath150
, there are 7 bs stars and 5 bhbs for a ratio of 1.4:1 . in the northern clump , an estimate of 450 bss is made by assuming a background similar to that in the upper left hand panel of [ fg13 ] , giving a ratio of 2.2:1 when combined with the 200 bhbs seen .
the bss are near the limit of detection in the north , and the sample is incomplete , so the ratio may be considerably larger . for the southern structure ,
about 380 bss are estimated , compared with about 60 bhbs , for a ratio of 6:1 , with considerable uncertainty . for a - colored stars in the field
not associated with the clumps , a ratio of about 2:1 is estimated by the same technique .
@xcite explore the populations of blue stars located at @xmath151 kpc and find a relatively large specific ratio of blue metal poor to field horizontal branch objects ( which are similar to the bs and bhb categories characterized here , though based on somewhat different selection criteria ) .
this apparent variation in the bs : bhb ratio between globular clusters , in the field and in the northern and southern structures , while still somewhat uncertain , appears to be a significant effect , indicating different populations of stars populate different regions of the halo . the large angular extent and
relatively low surface brightness of these structures leads one to believe that these may be tidally disrupted streamers in the halo .
the fact that two such structures were detected in 370 square degrees of data ( @xmath152% of the sky ) leads one to ask how common these structures might be in the galactic halo .
if the distribution of these objects is uniform across the sky , there may be as many as 200 of these to @xmath153 magnitude in @xmath36 .
since the size of the structures parallel to the line of sight could be as large as 10 kiloparsecs , the area effectively surveyed may be up to five times larger than the survey area ( since the width of the stripe is approximately 2 kpc at @xmath154 magnitude ) . in this case the total expected number of such structures in the halo would be 40 , or fewer if the length of each streamer is much greater than 20 kpc .
the existence of such structures lends strength to the argument that some or all of the halo of our galaxy has been formed through the accretion , or at least the disruption , of smaller stellar systems @xcite .
numerous studies have found evidence for coherent structures in the halo using kinematic information .
@xcite found a clump of 69 halo stars about 45 kpc above the north galactic pole with very similar radial velocities and proper motions and identified a moving group . @xcite
simulate satellite galaxy disruption and phase - space diffusion in the halo and suggest that the halo could consist of 300500 streams with components that pass close to the sun , a number consistent with that estimated above .
earlier , @xcite noted a group of 7 field horizontal branch stars with similar radial velocities all within 1 square degree on the sky and furthermore noted a positive signal in the 2-point star - star correlation function of a - colored objects on scales of 10@xmath155 . a cross - check against 21 cm emitting high velocity clouds @xcite does not show any obvious positional matches with these large structures in the north and south , but if even small amounts of gas is associated with the structures , a scenario such as that described in @xcite could hold , where the halos of galaxies are full of small complex , quasar absorption line producing structures .
if the structures are as numerous as 100 in the halo , then they could contain a few@xmath156 , a significant fraction of the luminous halo .
this is corroborated by figure 13 , from which one estimates that one in four of the bhb stars in the combined samples are in one of the two large clumps ( using the color separation of stellar types for field stars , and magnitude separation of the stellar types in the clumps ) .
to review : for the structure in the north , one sees bhb and bs candidate stars and for the structure in the south , which is closer , one sees these related populations of stars : bhb candidates , bs candidates , and f stars in numerical proportions which are not completely out of line with expectations from globular clusters or that of known galactic populations .
the rough mass estimates for the structures are somewhat above that for a massive globular cluster , and only a factor of a few smaller than that of known dwarf companions to the milky way .
thus galaxies like ursa minor or the draco dwarf spheroidal at about 70 kpc from the galactic center , which currently have central stellar densities much higher than those inferred for the extended structures in the north and the south , could provide progenitors of the low density structures seen in figure [ fg3 ] .
it should be noted that the mass estimates from bhb stars is highly uncertain , and assumes similar horizontal branch structure for all populations .
if bs counts were used , the mass estimates would be several times higher .
the discovery of extensive structures in the outer galactic halo is extremely interesting , and the number of a - colored stars associated with structures is of the same order as that in a supposed smooth background distribution of halo tracers .
this renders attempts to fit a smooth spheroid model with a single values of @xmath101 and @xmath18 problematic .
the approximate separation of bhb and bs stars in [ split ] , however , can be used to try and isolate only bhb tracers , and if one restricts that sample further to avoid the obvious clumps , one may still attempt a significant fit to a smooth spheroid .
figures [ fg20 ] shows a maximum likelihood fit to flattened spheroidal models of bhb candidate stars in the north and south .
fits were restricted to those bhb color candidates with implied @xmath157 , to regions which avoided the clump at @xmath158 in the south , and to regions which avoided saturated detections of stars ( which occurs in the sdss at about @xmath159 ) . constraining the fit to match
the numbers of objects seen in the north and south simultaneously helps greatly in restricting parameter space .
the model fit uses @xmath160 , where @xmath161 , and @xmath162 are galactic rectangular coordinates . setting @xmath163 , @xmath101 and @xmath164
are allowed to vary .
the maximum likelihood estimate of the halo structure parameters for @xmath165 is : @xmath166 , in agreement with @xcite and many other authors .
models with @xmath132 or @xmath167 yield significantly worse fits to the sample . after the numerous cuts placed on the input candidate sample , the number of stars fit is 527 stars .
the fit can be extrapolated to provide a local density for bhb stars at a distance of 8 kpc from the galactic center .
this normalization yields @xmath168 , in reasonable agreement with earlier studies @xcite .
more data , and cleaner spectroscopic separation of bhb from bs stars will of course help improve the confidence in these numbers to @xmath169 kpc , though if the presence of enormous structures in the outer halo at larger radii is pervasive , fitting a single set of parameters to the shape of the outer halo may not provide much insight into its overall mass distribution .
* a sample of 4208 a - colored stars from the sdss is uniform to @xmath62 and extends to @xmath56 .
this sample is used to probe halo structure at distances of up to @xmath170 kpc from the galactic center . *
a plot of all a - colored stars in magnitude - angle space shows significant clumpiness , including a large structure in the north at ( @xmath101 , @xmath96 ) = ( 220@xmath1 , @xmath171 ) extending over 30 degrees in length and at least 2.5 degrees in width .
the remarkable detection of a clump of 80 rr lyraes in this same area of sky by @xcite demonstrates the coherence and existence of this structure . in the south ,
a second extended structure of over 20 degrees in length is apparent .
numerous smaller structures , including the known dwarf irregular sextans ( at 75 kpc from the galactic center ) and the globular cluster pal 5 are also clearly visible . *
the example of pal 5 , which has a bhb and a bs sequence two magnitudes below its bhb , suggests an explanation for the parallel magnitude - angle - space arcs seen in the larger blue star halo samples .
photometric separation of a star surface gravities supports this explanation .
a majority of the high surface gravity a - colored stars in the halo may be bss . *
based on followup spectroscopy of eight objects , there are at least two sub - populations of a - colored stars identified : blue horizontal branch stars with low surface gravities , and blue straggler stars with high surface gravities and absolute magnitudes one to three magnitudes fainter than the horizontal branch . * the assumption of @xmath108 for the horizontal branch yields approximate distances to the large clumps in the north and south of 46 kpc and 33 kpc from the galactic center at positions ( x , y , z ) = ( 21 , -5 , 38 ) and ( -22 , 6 , -24 ) kpc , respectively .
the inferred masses of the detected structures is at least @xmath147 in the north and at least @xmath172 in the south , though it is not known what fraction of the total extent of these elongated structures has so far been detected . from estimates of the mass in these structures and the fraction of sample bhb stars they contain , one concludes that structures such as these could contain a significant fraction of the mass of the luminous halo . * fitting the ` field ' bhb candidate stars , ( those which avoid the structures ) , while requiring a consistent normalization between northern and southern galactic samples , yields a maximum likelihood fit to a spheroidal halo model of @xmath173 and flattening of @xmath8 for @xmath174 kpc , consistent with previous work .
since clumps of a - colored objects dominate the numbers of halo bhbs and bss at distances @xmath175 kpc , however , simple parameter fits to single spheroid models are not representative of the whole picture .
the existence of enormous structures in the outer halo must be addressed .
we acknowledge useful discussions with r. kron , b. gibson and many important suggestions and comments from the referee , t. beers .
we thank r. mcmillan and k. gloria for excellent assistance in operating the apo 3.5 m telescope .
the sloan digital sky survey ( sdss ) is a joint project of the university of chicago , fermilab , the institute for advanced study , the japan participation group , the johns hopkins university , the max - planck - institute for astronomy , princeton university , the united states naval observatory , and the university of washington .
apache point observatory , site of the sdss , is operated by the astrophysical research consortium .
funding for the project has been provided by the alfred p. sloan foundation , the sdss member institutions , the national aeronautics and space administration , the national science foundation , the u.s .
department of energy , and monbusho .
the sdss web site is http://www.sdss.org/. beers , r. c. , preston , g. w. , shectman , s. a. , doinidis , s. p. , and griffin , k. e. 1992 , , 103 , 167 chiba , m. , and beers , t. c. 2000 , , in press , astro - ph 0003087 djorgovski , s. g. , & meylan , g. in `` structure and dynamics of globular clusters '' , eds .
djorgovski , s. g. , & meylan , g. 1993 , asp conf series 50 , 370 .
doinidis & beers 1989 , , 340 , l57 eggen , o. j. , lynden - bell , d. , and sandage , a. 1962 , , 136 , 748 fan , x. , et .
2000 , , 119 , 1 fukugita , m. , ichikawa , t . , gunn , j. e. , doi , m. , shimasaku , k. , schneider , d. p. 1996
, , 111 , 1758 gliese , w. & jahreiss , h. 1979 , a&as , 38 , 423 gould a. , flynn , c. & bahcall , j. n. , 1998 , , 503 , 798 gunn , j. e. et al .
1998 , , 116 , 3040 harris , w. e. 1976 , , 81 , 1095 helmi , a. , & white , s. d. m. 1999 , 307 , 495 helmi , a. , white , s. d. m. , de zeeuw , p. t. , and zhao , h. , 1999 , nature , 402 , 53 houk , n. , & smith - moore , m .
1988 `` michigan catalogue of two - dimensional spectral types for the hd stars , vol .
4 '' univ of michigan .
ibata , r. a. , gilmore , g. , and irwin , m. j. 1995 , , 277 , 781 ivezic et al .
2000 , , submitted .
johnston , k. v. 1998 , , 495 , 297 johnston , k. v. , majewski , s. r. , siegel , m. h. , reid , i. n. , & kunkel , w. e. 1999 , , 118 , 1719 kinman , t. d. , suntzeff , n. b. & kraft , r. p. 1994
, aj 108 , 1722 kurucz , r. l. , in precision photometry : astrophysics of the galaxy , ed .
a. g. d philip , a. r. upgren , & k. a. janes 1991 ( schenectady : davis ) , 27 layden , a. , hanson , r. , hawley , s. l. , klemola , a. r. , and hanley , c. j. 1996 , , 112 , 2110 lenz , d. d. , newberg , h. j. , rosner , r. , richards , g. t. , and stoughton , c. 1998 , , 119 , 121 lupton , r. h. , gunn , j. e. , & szalay , a. s. 1999 , , 118 , 1406 lynden - bell , d. , & lynden - bell , r. m. 1995 , , 275 , 429 majewski , s. r. 1994 , , 431 , l17 majewski , s. r. 1999 , in the third stromlo symposium : the galactic halo , eds . b. k. , gibson , t .
s. axelrod , and m. e. putman , asp conf .
ser . , 165 , 76 majewski , s. r. , munn , j. a. , & hawley , s. 1994 , , 427 , l37 newberg , h. j. , richards , g. t.,richmond , m. , and fan , x. 1999 , , 123 , 377 norris , j. e. , and hawkins , m. r. s.1991 , , 380 , 104 pier , j. r. 1983 , , 53 , 791 preston , g. w. , beers , t. c. , & shectman , s. a. 1994 , , 108 , 538 preston , g. w. , shectman , s. a. , & beers , t. c. 1991 , , 375 , 121 rodgers , a. w. , and roberts , w. h. 1993 , , 106 , 2294 rodgers , a. w. , harding , p. , and sadler , e. 1981 , , 244 , 912 saha , a. 1985 , , 289 , 310 schlegel , d. j. , finkbeiner , d. p. , davis , m. 1998 , , 500 , 525 sills , a. , and bailyn , c. d. 1999 , , 513 , 428 sluis , a. , & arnold , r. 1998 mnras , 297 , 732 sommer - larsen , j. , christensen , p. r. , and carter , d. 1989 , , 238 , 225 stryker , l. l. 1993 , , 105 , 1081 van dyk , s. d. , puche , d. , & wong , t. 1998 , aj 116 , 2341 wakker , b. p. & van woerden , h. 1997 , ara&a 35 , 217 wetterer , c. j. & macgraw , j. t. 1996 , aj 112 , 1046 wilhelm , r. , beers , t. c. , sommer - larsen , j. , pier , j. r. , layden , a. , c. , flynn , c. , rossi , s. , christensen , p. r. 1999 , , 117 , 2329 york , d. g. , dopita , m. , green , r. , bechtold , j. 1986 apj 311 , 610 york , d.g .
2000 , aj , submitted .
zinn , r. , 1985 apj 293 , 424 lllllrrr + & & & & & & & + cs29516@xmath1760011@xmath177&12.24@xmath178 & 1.30@xmath178 & -0.17@xmath178&21&71 & -42 & bhb + hd13433 & @xmath1798.16@xmath178&1.10@xmath178 & -0.14@xmath178 & 35 & 184 & -68 & a0v + wd0148p467@xmath180 & 12.44@xmath178&0.20@xmath178 & -0.17@xmath178 & 60 & 134 & -15 & da wd + sdssp [email protected]&17.20&1.13 & -0.20&25 & 1 & 46 & bhb@xmath181 + sdssp [email protected]&18.88&1.12 & -0.13 & 36 & 1 & 46 & bs@xmath181 + sdssp [email protected] & 17.55 & 0.99 & -0.29&31 & 6 & 40 & bhb + sdssp [email protected] & 17.52 & 1.18 & -0.24 & 26 & 14 & 32 & bhb + sdssp [email protected] & 17.49 & 1.18 & -0.12 & 27 & 11 & 37 & am + sdssp [email protected] & 17.49 & 0.96 & -0.26 & 33 & 15&32 & a / bs + sdssp [email protected] & 17.82 & 1.11&-0.23 & 26 & 13&33 & bhb + sdssp [email protected] & 17.54 & 1.11&-0.20 & 26 & 13&31 & bhb + | a sample of 4208 objects with magnitude @xmath0 and colors of main sequence a stars have been selected from 370 square degrees of sloan digital sky survey ( sdss ) commissioning observations . the data is from two long , narrow stripes , each with an opening angle of greater than 60@xmath1 , at galactic latitudes @xmath2 on the celestial equator .
relative photometric calibrations good to 2% and consistent absolute photometry allows this uniform sample to be treated statistically over the large area .
an examination of the sample s distribution shows that these stars trace considerable substructure in the halo .
large overdensities of a - colored stars in the north at @xmath3 and in the south at @xmath4 and extending over tens of degrees are present in the halo of the milky way .
@xcite has detected the northern structure from a sample of rr lyraes in the sdss . using photometry to separate the stars by surface gravity ,
both structures are shown to contain a sequence of low surface gravity stars consistent with identification as a blue horizontal branch ( bhb ) .
both structures also contain a population of high surface gravity stars two magnitudes fainter than the bhb stars , consistent with their identification as blue stragglers ( bss ) .
the majority of the high surface gravity stars in the galactic halo may be blue straggler stars like these .
a population of f stars associated with the a star excess in the southern structure is detected ( the f stars in the northern structure at 46 kpc would be too faint for the sdss to detect ) . from the numbers of detected bhb stars , lower limits to the implied mass of the structures are @xmath5 and @xmath6 , though one does not yet know the full spatial extent of the structures . the fact that two such large clumps have been detected in a survey of only 1% of the sky indicates that such structures are not uncommon in the halo .
simple spheroidal parameters are fit to a complete sample of the remaining unclumped bhb stars and yield ( at @xmath7 kpc ) a fit to a halo distribution with flattening ( @xmath8 ) and a density falloff exponent of @xmath9 . |
in recent years , doped manganese oxides remained at the forefront of theoretical and experimental research@xcite .
the main source of interest in theses systems is the phenomenon of colossal magnetoresistance ( cmr ) , which they exhibit , and that is likely to have important technological applications . in the meantime ,
the underlying basic physics remains elusive , and probably involves the strongly - correlated nature of the doped magnetic oxides .
the cmr in doped manganates is observed for intermediate hole - doping levels , typically @xmath0 , in the temperature region around the transition between low - temperature metallic ferromagnetic ( fm ) and high - temperature insulating paramagnetic phases .
in addition to double - exchange ferromagnetism@xcite , the cmr compounds also possess pronounced antiferromagnetic ( afm ) tendencies , as evident from the afm spin ordering with nel temperatures of about 100 - 200k , observed@xcite at the doping end - points ( @xmath1 and @xmath2 ) .
this antiferromagnetism is of a superexchange origin@xcite .
the manganates physics involves several degrees of freedom of substantially different nature , including localized core spins @xmath3 of mn ions , fermionic degrees of freedom associated with conduction @xmath4-electrons , lattice distortions , _
etc_. in such systems , the presence of competing interactions ( such as fm and afm ) often gives rise to phase separation@xcite , whereby areas of different phases are stabilized in a structurally and stoichiometrically homogeneous sample . in the case of the manganates
, it has even been suggested@xcite that phase separation into insulating paramagnetic and metallic fm phases may explain the resistivity peak observed near the curie temperature .
in the present paper , we focus on the low - temperature regime where the presence of phase separation in the appropriate manganate systems has been directly verified , _
e.g. , _ by means of scanning tunneling microscopy ( stm ) on thin films@xcite .
transport measurements reveal metastability and history dependence near the percolation threshold ( @xmath5 ) , confirming phase separation in both film@xcite and crystalline@xcite samples . using simple microscopic models@xcite it can readily be shown@xcite that the hole concentration @xmath6 indeed controls the balance between the fm and afm tendencies of the system . once @xmath6 is tuned away from the optimal cmr doping region , the homogeneous fm metallic state no longer corresponds to the energy minimum . instead , energy can be gained by changing the magnetic ordering , carrier density , bandstructure , and/or orbital state in _ part _ of the system , making the sample inhomogeneous@xcite .
the surface tension between different phases@xcite then competes against the long - range interactions present in the system in the form of electrostatic forces@xcite or long - range crystal strain fields@xcite .
these require that the system remains homogeneous at least _ on average _ on the appropriate length scale ( such as the debye
hckel screening length ) , resulting in a periodic arrangement of nano- or mesoscopic regions of different phases@xcite .
the geometry of the ensuing inhomogeneous ( phase separated ) state is at the focus of our present study .
early studies of phase separation in double - exchange
superexchange systems@xcite implicitly assumed that the effects of the discrete lattice are unimportant , and consequently treated the problem within the continuum based , long - wavelength , approach . within this framework
, the optimal phase separation geometry at small values of the fm volume ( or area ) fraction @xmath7 ( also the average magnetization per site ) is obviously that of spherical ( in two dimensions , circular ) fm droplets located at the sites of a packed hexagonal ( triangular ) super - lattice . to the best of our knowledge ,
only the three - dimensional case was treated in detail , with the implication that in two dimensions the situation is similar .
when the system parameters are varied in such a way that @xmath7 increases beyond @xmath8 , the geometry changes to that of spherical afm droplets in an otherwise fm matrix .
the change generally occurs via a direct `` geometrical phase transition''@xcite without any intervening regime characterized by both phases forming infinite connected shapes ( such as filaments and planar slabs in three dimensions or stripes in two dimensions)@xcite .
this latter conclusion is important , since such slab or stripe arrangements , if realized , would have been characterized by peculiar and potentially useful properties such as history - dependent anisotropy of the ground state resistivity . however , the continuum treatment , which is at the basis of this result , is not valid beyond the region of very small values of @xmath9 .
indeed , recent studies suggest@xcite that the boundary between the two phases is abrupt on the lattice - spacing scale ( _ i.e. , _ of the type commonly associated with ising spin systems ) . such a boundary can not be adequately described in the continuum limit , and its surface tension depends on its orientation with respect to the crystalline axes@xcite .
this directional dependence of boundary energies should in turn affect the droplet shape ( generally favoring diamond - shaped droplets in two dimensions)@xcite , the arrangement of droplets in space , and ultimately the way the geometry of phase separation evolves with varying @xmath7 .
this is apparently a generic property of electronic phase separation , found also within the frameworks of falikov
kimball@xcite and @xmath10 ( ref . )
models . in the present paper , we revisit the problem within the framework of a single - orbital double - exchange superexchange hamiltonian ( with infinite hund s coupling @xmath11 ) , augmented by a long - range coulomb interaction term . using a variational hartree - fock approach , we compute the energies of various two - dimensional droplet and stripe phases corresponding to a fm area fraction @xmath12 , and determine the optimal configuration .
our most important finding is that while a droplet lattice exists at low doping levels , a striped arrangement has a lower energy and is therefore stabilized over a broad region of the phase diagram . as anticipated from our previous results concerning the orientational dependence of the boundary energy@xcite , we find that diamond droplets and diagonal stripes are the preferred geometries for the fm regions of the inhomogeneous states .
these conclusions gain further support from unrestricted hartree - fock calculations which we have carried out using monte - carlo simulated annealing on moderate size clusters .
the simulations also demonstrate the existence of inhomogeneous states comprised of afm droplets ( or stripes ) embedded in a fm background ( @xmath13 ) , at higher doping levels .
while our results pertain to the two - dimensional case , it is likely that qualitatively our conclusions would also apply to three - dimensional systems .
specifically , we suggest that a phase separated state with filament or slab geometry ( rather than a lattice of droplets ) is realized for a certain range of parameters in three dimensions . in addition
, we find that the typical droplet size and stripe width do not exceed several lattice constants .
this means that the motion of the charge carriers is strongly quantized , rendering droplets midway between metallic bulk and magnetic polarons@xcite and giving rise to singularities in the stripe energy associated with the quantisation of the transverse kinetic energy .
this important property was not included in the earlier work@xcite , which assumed sufficiently large length scales for such quantum effects to be negligible .
our approach , on the other hands , allows one to explore the crossover between the regime of singly - occupied magnetic polarons , which appear for strong coulomb and afm couplings , and the more conventional phase separation behavior where each metallic droplet is populated by many charge carriers .
the paper is organized as follows : in sec .
[ sec : model ] , we introduce the model and briefly review the physics underlying phase separation and magnetic polaron formation in the absence of a long - range force . a brief description of the calculational methods which were implemented in order to include the effects of the long - range coulomb repulsion appears in sec . [ sec : methods ] , while the mass of details is relegated to the appendices .
[ sec : results ] contains a detailed description of our hartree fock and monte - carlo results .
we conclude with a brief discussion of the results in the context of current experimental and theoretical work ( sec . [
sec : conclu ] ) . while an arrangement of conducting and insulating stripes in doped manganate films has not yet been observed , we suggest that present experimental knowledge should allow for a meaningful and successful research effort in this direction .
the starting point for the following calculation is the two - dimensional double - exchange hamiltonian , generalized to include the superexchange coupling and the long - range coulomb interaction , @xmath14 here @xmath15 is the nearest - neighbor hopping amplitude and @xmath16 annihilates a conduction electron of spin @xmath17 at site @xmath18 of a square lattice .
@xmath19 denotes the core spin made of three d - shell electrons ( @xmath20 ) localized at site @xmath21 , whose afm superexchange interaction with neighboring core spins is given by the second term in @xmath22 .
the third term arises from hund s coupling between the core spins and the conduction electrons , where the spin operator for the conduction electrons on site @xmath21 involves the pauli matrices @xmath23 .
it is this term , in conjunction with the fact that the hopping preserves the electronic spin , which gives rise to the double - exchange mechanism .
this favors a fm spin configuration in order to reduce the conduction electrons kinetic energy@xcite .
the last term includes the coulomb interaction among the conduction electrons , whose average density is @xmath6 , and a neutralizing uniform positive background , created by the donors .
owing to the long - range nature of the coulomb interaction , the atomic - scale inhomogeneities of this background in real systems ( created by chemical substitution ) are not expected to be important from the point of view of our main purpose of comparing the energies of various inhomogeneous phases .
this is because such energies always involve integration over volume . in using the simplified model ,
( [ eq : ham_orig ] ) , we neglect some additional physics characteristic of the cmr manganates@xcite .
this includes the presence of two ( rather than one ) conduction electron @xmath4-bands and the electron - lattice coupling .
the logics behind this simplification is summarized , _
e.g. , _ in ref .
: it is assumed that the mechanism for phase separation ( charge ordering ) is the competition between ferro- and antiferromagnetism in the presence of a long - range coulomb repulsion [ all contained in eq .
( [ eq : ham_orig ] ) ] .
once the charge ordering is established , in a real system the orbital ordering ( and the lattice distortions ) would follow , leading to a quantitative renormalization of the parameter values . the model , eq .
( [ eq : ham_orig ] ) , is however expected to suffice for a qualitative study of the generic features of the phase diagram while its simplicity allows to maintain clarity of analysis .
further arguments regarding the expected model - independence of our conclusions shall be given in sec .
[ sec : conclu ] .
the relatively large value @xmath24 of the mn core spins means that their fluctuations are small , particularly in the @xmath25 limit considered here . in the following we assume @xmath26 and treat the core spins classically .
consequently , the effective hamiltonian governing the physics of the conduction electrons is determined by the configuration of the classical spins @xmath27 .
as far as the hund s coupling is concerned the manganates are characterized by a moderate bare @xmath28 .
however , they also include a strong hubbard on - site repulsion , @xmath29 , which significantly renormalizes @xmath11 towards the strong coupling limit@xcite .
therefore , while we omit the hubbard interaction from our hamiltonian ( [ eq : ham_orig ] ) , we model its effects by taking @xmath30 . band theory calculations@xcite suggest that typical values of the hopping amplitude @xmath15 in the cmr manganates lie between 0.3 ev and 0.5 ev . the value of @xmath31 can be roughly estimated from the
experimentally observed nel temperatures in the fully doped or undoped ( with no conduction @xmath4 electrons or with no holes ) case@xcite , @xmath32 , corresponding to @xmath33 .
the long - range coulomb interaction strength , @xmath34 , for thin films is evaluated as @xmath35 . here ,
@xmath36 is the electron charge and @xmath37 is the lattice spacing . the _ effective _ dielectric constant @xmath38 is given by the average of dielectric constant @xmath39 of the substrate and that of the air , @xmath40 .
among the substances which can be used as substrates for manganate films , lanthanum aluminate and neodymium gallate have@xcite @xmath41 and @xmath42 , yielding @xmath43 ev and @xmath44 ev respectively .
dielectric properties of the third possible substrate , strontium titanate , are strongly dependent on temperature , with@xcite @xmath39 changing from 24000 ( corresponding to @xmath45 mev ) at 4.2k to @xmath46 ( @xmath47 ev ) at 300k .
this suggests a possibility of experimentally varying the value of @xmath34 by using different substrates and/or changing temperature .
theoretical investigations of the double - exchange superexchange competition have a history of more than 40 years .
it is by now firmly established@xcite that this competition is resolved not via a second - order phase transition from the fm state to a uniform state with a helical or canted magnetic ordering , but rather via separation of the sample into regions characterized by different spin arrangements and conduction electron bandstructures
. we will be interested in the case of phase separation into fm and afm regions with abrupt ising - type boundaries@xcite between them .
this means that the resulting configuration of @xmath19 remains collinear , with all core spins either parallel or anti - parallel to a selected direction .
thus , it is possible to denote a spin state simply by @xmath48 , and the hamiltonian of the conduction electrons becomes a function of @xmath49 .
the large hund s exchange coupling then forces the conduction electrons spins to polarize in parallel with the core spins , resulting@xcite in the following distribution of hopping amplitudes for a given spin configuration @xmath50 : @xmath51 after these simplifications , the hamiltonian takes the form @xmath52 the model , eq .
( [ eq : ham_simp ] ) , on the ( bi - partite ) square lattice is invariant under the particle - hole transformation @xmath53 , and @xmath54 , where @xmath55 takes the values 0,1 on the two sublattices . as a result
we note that in the following , @xmath6 acquire the more general meaning of a carrier density , i.e. , either the electronic density or the hole density relative to the half - filled state .
we will now briefly review the ground state properties of the hamiltonian ( [ eq : ham_simp ] ) at @xmath56 . when the carrier density @xmath6 is finite , the ground state of the system at @xmath57 is uniform and fm . with increasing @xmath31 beyond a certain critical value @xmath58 ,
this uniform fm state eventually becomes destabilized , and a non - uniform ground state is obtained instead . in this _ phase separated _ state only part of the system is occupied by the fm phase .
we will be interested in the case in which the other part is a simple nel antiferromagnet with zero charge - carrier density .
a variational study@xcite shows that in two dimensions such a phase separated state may be realized only for @xmath59 . at higher values of @xmath31 the magnetic
ordering in either the electron - rich or the electron - poor regions of the sample differs from that of a ferromagnet or a nel antiferromagnet . in the fm region of a macroscopically phase separated state . ] thermodynamic equilibrium between macroscopic fm and afm regions means that the thermodynamic potentials in the two phases are equal , @xmath60 where @xmath61 and @xmath62 @xmath63 is the density of conduction electron states in the fm region .
since the latter is large , @xmath63 can be taken to be the two - dimensional tight - binding density of states , and boundary corrections may be neglected . by solving eqs .
( [ eq : tdstability][eq : omegaafm ] ) for the fermi energy , @xmath64 , one readily obtains the carrier density in the fm region , @xmath65 , via @xmath66 with the result depicted in fig .
[ fig : xfmtd ] .
the system remains in a uniform fm state as long as @xmath67 .
the fraction of the system area ( or volume ) , occupied by the fm phase , is given by @xmath68 .
the critical value , @xmath69 , for the onset of phase separation is then determined by the condition @xmath70 .
in addition to the macroscopic phase separation as described above , the double exchange - superexchange competition can also be resolved via an altogether different scenario ( formation of magnetic polarons ) . when a single electron ( or hole ) is lodged into an antiferromagnetically ordered double - exchange magnet with zero charge - carriers , ( @xmath1 ) , a free ( self - trapped ) magnetic polaron , or ferron@xcite ,
is formed around it .
it is essentially a microscopic fm region , containing one charge carrier , in an otherwise afm system . since the propagation of charge is unimpeded in the fm region it acts as a potential well for the sole carrier , which occupies the lowest bound state inside the well .
the polaron binding energy , @xmath71 , ( with respect to the state where the afm order is unperturbed ) can be be easily estimated@xcite .
we consider the case of a diamond - shaped polaron , with @xmath72 sites along each side ( see fig .
[ fig : geometry ] , upper left ) .
for @xmath73 we find @xmath74 where the first two terms are the ground - state energy of the charge carrier and
the last one represents the superexchange contribution .
expression ( [ polenergy ] ) should be minimized with respect to @xmath75 , resulting in @xmath76 here , the coefficient of the second term depends on the geometry of the fm micro - region ( _ e.g. _ , for a round polaron one would have obtained @xmath77 instead of @xmath78 ) . the above expressions are valid in the @xmath79 regime ( yielding @xmath80 ) , where it is easy to verify an important statement which is expected to hold for all @xmath31 .
namely , if in the absence of a coulomb interaction @xmath34 , a second carrier is added to the system , it is energetically favorable for the two charge carriers to occupy the two lowest bound states in a _ shared _ fm micro - region rather than to form two independent polarons .
this conclusion is verified by calculating the binding energy of the ( diamond - shaped ) doubly - occupied polaron @xmath81 which clearly satisfies @xmath82 .
this trend continues when further charge is added , and at @xmath83 the binding energy ( per carrier ) of the @xmath84-carrier polaron decreases toward the limiting value @xmath85 , @xmath86 which is the energy gain per carrier associated with the macroscopic phase separated state .
the latter can be evaluated as @xmath87 where @xmath88 and @xmath89 are the energies per site of the fm and afm phases .
using eqs .
( [ eq : tdstability ] ) and ( [ eq : tdxfm ] ) to evaluate @xmath64 and @xmath65 we find , in the limit @xmath90 , @xmath91 the inequality , @xmath92 , implies that for any finite carrier density , at @xmath56 , the double - exchange superexchange competition is resolved via macroscopic phase separation . notwithstanding the preceding discussion , its conclusion may change if a realistically strong coulomb interaction @xmath34 is included , favoring a large spatial separation between the charge carriers . indeed , as we demonstrate in the following , a polaronic state arises in the regime of large @xmath34 and small carrier density .
it is the extreme limit of a broad range of inhomogeneous states which originate from the frustration of macroscopic phase separation by long - range forces .
the study of this intermediate region of parameters lies at the focus of the remaining part of the paper .
since the typical size of the resulting fm regions is rather small one needs to take into account the effects of quantization of the charge carrier motion . at the same time , some of the results obtained for the macroscopic phase - separated system , such as the directional dependence of the boundary energy@xcite , still offer important guidance to the understanding of the inhomogeneous configurations .
next , we outline the methods used to treat this intermediate regime which is characterized by a combination of both traditional phase separation and magnetic polaron ( quantized ) physics .
given the hamiltonian ( [ eq : ham_simp ] ) , our task is to find the configuration of core spins in the ground state .
however , there is a vast multitude of possible spin configurations amongst which the ground state is to be sought , making it impossible to explore all of them .
nevertheless , previous studies of similar or related systems@xcite suggest several families of highly symmetrical configurations as reasonable ground state candidates .
the two main types of spin configurations studied in this work are fm droplets in an afm background , and alternating fm and afm stripes , as illustrated in fig .
[ fig : geometry ] . a uniform fm phase , in which the double - exchange mechanism completely overcomes the superexchange , is also considered .
[ cols="^,^",options="header " , ] calculating the energy of the conduction electrons in a given configuration of core spins is a difficult problem .
here we suffice with the hartree - fock ( hf ) approximation , which gives an upper - bound to their ground state energy .
since we are dealing with periodic spin configurations , the hf equations for the whole system can be rewritten as an effective eigenvalue problem within a single unit cell . the superexchange contribution to a configuration s energy
is simply calculated by counting the number of fm and afm bonds in a unit cell .
based on previous analytical results@xcite and numerical investigations@xcite the considered droplets are either diamond or square shaped , and are chosen to form either a triangular or square super - lattice ( see fig .
[ fig : geometry ] ) .
several droplet phases are possible by combining different droplet shapes and super - lattice types . in addition
, one has variational freedom to specify @xmath75 , the size of the fm droplets , and @xmath84 , the number of conduction electrons in each one of the droplets .
the distance between the droplets , @xmath93 , is then uniquely determined by the type of super - lattice , by @xmath84 , and by the average density of conduction electrons , @xmath6 .
the energy of each droplet phase is found by minimizing its energy density , @xmath94 with respect to the variational parameters , @xmath75 , and @xmath84 .
here @xmath95 is the hf energy of the conduction electrons inside a unit cell containing a single droplet , and @xmath96 is the afm energy _ per site_. the details of the hf calculation appear in appendix [ app : hartree ] .
the main source of complication is the necessity to take into account the hartree interaction between electrons belonging to different droplets in the infinite super - lattice .
this is done by employing ewald s summation method ( see appendix [ app : ewald ] ) .
the afm coupling energy per lattice site is @xmath97 where @xmath98 is the area of a unit cell .
two types of stripe phases were considered : diagonal and bond - aligned .
additional variational freedom comes from the need to specify @xmath99 , the stripe width , and @xmath65 , the ( average ) density of conduction electrons within the fm stripe . just as for the droplet phase ,
the stripe phase energy is found by minimizing its energy density @xmath100 with respect to @xmath99 and @xmath65 .
here @xmath101 and @xmath102 are , respectively , the conduction electrons number and energy per unit cell , and @xmath103 is the afm energy per site .
a unit cell in diagonal stripes is only one lattice spacing long in the direction along the stripe , and two spacings long in bond - aligned stripes ( see fig .
[ fig : geometry ] ) ; its width equals the stripe periodicity .
therefore , @xmath104 @xmath101 , together with @xmath6 , uniquely determine the distance between stripes @xmath105 . in a similar manner to the case of the droplet phase , when calculating the hf energy one needs to take into account the hartree interaction between the infinite number of unit cells in the systems .
moreover , the extended nature of the states along the stripes means that it is necessary to consider also the fock exchange between different unit cells on the same stripe .
a detailed account of the way this is done is presented in appendix [ app : hartree ] .
the afm spin coupling energy per unit area for both diagonal and bond - aligned stripe phases is @xmath106,\ ] ] where @xmath68 is the fraction of fm regions in the system . by comparing the energies of all the above mentioned phases ,
a phase diagram is constructed , depicting the nature of the ground state as a function of the external parameters , @xmath6 , @xmath107 and @xmath108 .
we have supplemented the calculation of the hf energy of various variational configurations by an _
unrestricted _ hf calculation using monte - carlo simulated annealing . in this method , the energy of a finite - sized system is minimized with respect to the full configuration space of core spins , rather than a special subset of spin textures . in each monte - carlo step the energy of a given configuration of classical core spins is evaluated using the hf approximation .
a spin configuration is accepted as the system s new state if the change in energy from the current state satisfies the metropolis condition .
the temperature is slowly decreased until the system reaches a stable , low energy configuration . if the temperature is decreased slowly enough , the final state is the hf approximation of the ground state .
the underlying assumption of the present study is that the system indeed separates into fm and afm regions with an abrupt boundary between them .
therefore , the mc simulation needs to explore only such configurations , improving the convergence time .
this can be achieved by setting all the spins on one sub - lattice to the `` up '' state , and incrementally flipping the spins on the other sub - lattice .
an additional improvement comes from a new algorithm used to decide which spin to flip . at first
, a spin is chosen randomly .
it is flipped if the resulting state satisfies the metropolis criterion .
if the spin is located near a fm - afm boundary then its neighbors are added to a queue of spins to be tested for flipping .
after all the spins in the queue have been tested for a flip , a new spin is chosen randomly .
requiring that a spin be added to the queue no more than once , prevents the simulation from repeating itself , thus maintaining ergodicity .
in our calculations , the system contained @xmath109 sites arranged periodically on a torus .
a linear annealing schedule was employed over @xmath110 mc sweeps , and an identical number of sweeps at the lowest temperature allowed the system to thermalize into the ground state .
the temperatures started from above @xmath111 at the beginning of the annealing schedule to below @xmath112 at its end .
even though this method minimizes the system s energy with respect to an unrestricted configuration space , it has a number of disadvantages when compared to the variational hf method , applied to only a number of special configurations .
first , its periodicity is fixed ; in our case it is @xmath113 sites along each axis .
in addition , the presence of long range interactions causes the simulations to converge very slowly .
nevertheless , it provides an important reference point with which the the variational hf results may be contrasted , especially in order to confirm that the variational manifold contains the most relevant configurations .
in the present section , we present and analyze our numerical results . the coupling constants @xmath31 and
@xmath34 are measured in units of @xmath15 , by setting the hopping amplitude @xmath114 .
the hf energies of all the considered phases were calculated in the parameter range @xmath115 , @xmath116 , and for three values of @xmath34 , namely @xmath117 .
we chose to concentrate on this region in the @xmath118 plane for two reasons .
as already mentioned , a previous estimate@xcite sets @xmath119 as the upper limit for the realization of a fm nel afm ( as opposed to other types of magnetic ordering ) phase - separated state in two - dimensions .
secondly , our calculations indicate that the line @xmath120 , crosses @xmath121 at @xmath122 , see fig .
[ fig : phsdgm ] .
the region below this line in the @xmath118 plane corresponds to configurations in which the fm phase occupies more than half the system area .
while the stripes phases , which we consider , continue to be relevant in this region , we expect ( and confirm in our mc simulations ) that the phases of fm droplets ought to be replaced by configurations of afm droplets embedded in a fm background .
the latter turn out to be more involved computationally and were left out of the present study .
we also wish to note that the above values of @xmath123 and @xmath124 , are sensitive to the details of the considered model . therefore , while experimentally , percolation of the metallic phase at low temperatures is observed in manganates with @xmath125 , we expect our qualitative conclusions to apply to more complicated models of manganates , as long as phase separation into fm and afm phases is possible .
we begin our review of the results by discussing the phase diagram and presenting general arguments for its structure .
we then move on to consider the details of the most dominant phases . .
] the main result of our calculation is the phase diagram , fig .
[ fig : phsdgm ] , derived from the variational hf approach and depicting the system s ground state configuration as a function of the parameters , @xmath6 , @xmath31 and @xmath34 .
it demonstrates that diamond droplets in a triangular formation is the preferred phase at low densities , while diagonal stripes are prevalent at higher values of @xmath6 .
this stability of a striped phase is the most important result of our calculation .
the striped arrangement is expected to possess unusual and potentially useful properties ( see sec .
[ sec : conclu ] below , where we also mention possible directions of experimental search for the stripe phase in cmr manganates ) .
when the coulomb interaction strength @xmath34 is increased , the transition between droplets and stripes occurs at higher @xmath6 .
as discussed above , our variational approach becomes insufficient below the line @xmath120 , as we do not allow for a phase of afm droplets embedded in a fm background .
such a phase is expected to appear near the transition to the uniform fm state .
this conclusion is supported by the unconstrained hf results presented below .
we are unable , though , to map in detail the boundary between the stripe and droplet phases in this parameter regime .
the general features of the phase diagram can be explained by simple energy considerations .
the preference of diagonal stripes and diamond droplets is a direct result of the lower energy of diagonal boundaries , as previously established by two of the authors@xcite .
the appearance of a triangular droplet lattice at low densities is akin to the physics giving rise to the wigner crystal in a dilute gas of electrons .
next , we elaborate on the reasons and nature of the transition between the droplet and stripe phases . to this end , let us examine how the energy difference between the two phases evolves with @xmath6 . as @xmath6 increases , the distance between droplets or stripes diminishes , but our hf results indicate that the size of these fm regions and the electron density within them , @xmath65 , vary slowly in the vicinity of the phase transition .
the combined difference between the kinetic and magnetic energies per electron of the two phases , @xmath126 , depends on @xmath65 and on the size and shape of the fm regions , but not on the distance between them .
we therefore conclude that at the qualitative level , changes of @xmath126 with @xmath6 can not be the driving force behind the transition . instead , we concentrate on the doping dependence of the difference in the coulomb energy per electron between droplets and stripes , @xmath127
. the coulomb energy contains contributions coming from the interaction between charges within a single super - lattice unit cell and between different cells .
the neutrality of each unit cell ( due to the positive background ) implies that the dominant contribution to the coulomb energy per electron originates from the intra - cell component .
simple dimensional analysis allows us to obtain an estimate for its behavior .
the amount of positive background charge within a droplet unit cell is @xmath128 , @xmath93 being the inter - droplet spacing .
thus , the coulomb potential due to the positive background is @xmath129 .
the interaction between electrons within a droplet generates @xmath130 , where @xmath75 is the droplet size . since @xmath131
, we have @xmath132 the coulomb energy in the stripe phase takes a different form .
the amount of charge per unit length is @xmath133 where @xmath99 and @xmath105 are the stripe width and the distance between stripes , correspondingly .
the background potential is then @xmath134 and the potential due to electrons in the same stripe is @xmath135 .
together they give @xmath136 where in this case @xmath137 .
consequently , the difference in coulomb energy per electron between the droplet and stripe phases has the form @xmath138 , \label{dphi}\end{aligned}\ ] ] as a function of @xmath7 , demonstrating the difference between the two possible branches . ] and @xmath139 as a function of @xmath6 as obtained from the hf calculation for @xmath140 and various values of @xmath31 . below each plot
are colored bands showing the ground state configuration at the corresponding @xmath6 values : green ( light)- droplets , blue ( dark ) - stripes .
the transitions occur when @xmath141 . ] where @xmath142 and @xmath143 are numerical constants characterizing the geometry of droplets and stripes , respectively .
it is implicitly assumed in eq .
( [ dphi ] ) that in the transition region between the phases @xmath65 is the same for both configurations ( the hf calculation shows that this is correct up to @xmath144 ) .
the transition itself takes place at @xmath145 , satisfying @xmath146 where we have used the constancy of @xmath126 near the transition .
if @xmath147 then @xmath127 is a monotonously increasing function of @xmath7 ( in the physical range @xmath148 ) , see fig .
[ fig : dphith ] . in this case at most a single solution , @xmath145 , exists to condition ( [ eq : transition ] ) , implying that the droplet phase is preferred when @xmath149 , while stripes occur for @xmath150 ; the area of the droplet phase increases with @xmath34 . on the other hand ,
if @xmath151 , @xmath127 acquires a maximum and two solutions , @xmath152 and @xmath153 , may appear . under such conditions
a reentrant behavior follows , i.e. , droplets are preferred when @xmath154 _ or _
@xmath155 , and stripes are realized in the region @xmath156 , which grows with increasing @xmath34 .
we note that in any case the existence of a solution to condition ( [ eq : transition ] ) , crucially depends on the value of @xmath126 .
it is the latter which reflects the features taken into account for the first time in the present work ( _ viz .
_ , the orientational dependence of the boundary energy and the quantization of the carrier motion . ) fig .
[ fig : dphi ] shows hf results for @xmath127 and @xmath157 as a function of @xmath6 for various values of @xmath31 .
transitions between droplet and stripe phases occur when @xmath141 .
the division into discontinuous segments is due to changes in the properties of the stripes or droplets respectively ( the optimal values of @xmath99 , @xmath75 , and @xmath65 , see below ) .
however , the transitions generally do not occur at these points of discontinuity , leading to our previous assertions concerning the constancy of @xmath126 and the dominant role of the coulomb interaction in the vicinity of the transition .
the first two transitions at @xmath158 and @xmath159 are near a maximum in @xmath127 , demonstrating the @xmath151 branch behavior .
whereas these transitions occur on one continuous segment , the third transition occurs on a different segment , where only one solution exists .
other one solution transitions are shown for @xmath160 and @xmath161 . in general
, a triangular lattice of diamond shaped droplets proved to be energetically more favorable than the other types of droplet phases .
as noted before this is a consequence of the directional dependence of the fm - afm boundary energy and the minimization of the inter - droplet coulomb energy .
[ fig : ditr ] shows the optimal droplet size , @xmath75 , and the number of conduction electrons per droplet , @xmath84 , as deduced from the variational hf calculation . increasing the strength of the coulomb repulsion has the obvious effect of decreasing the droplet size .
specifically , for the case of @xmath162 the variational study yields very small ( @xmath163 ) singly occupied droplets in the regime of low @xmath6 and intermediate to large @xmath31 . comparing their energy to the other types of inhomogeneous states
reveals that these _ magnetic polarons _ are in fact the lowest energy configuration in this region of parameters , see the hf phase diagram , fig .
[ fig : phsdgm ] .
( top ) and number of electrons per droplet @xmath84 ( bottom ) . ]
fig.[fig : diastr ] shows the optimal stripe width @xmath99 and conduction electron density @xmath65 for diagonal stripes .
the latter are more favorable than their bond - aligned counterparts due to the orientation dependence of the boundary energy .
one striking feature in these hf results is the existence of abrupt transitions in the stripe width .
a small increase in @xmath6 may lead to a discontinuous change in @xmath99 . on the other hand , increasing @xmath31 typically leads to changes in @xmath99 which are less steep .
the electron density within the fm stripes , @xmath65 , varies , in general , very slowly with @xmath6 , and increases with @xmath31 .
( top ) and @xmath65 ( bottom ) .
the number of partially filled bands within a stripe changes across each black contour on the top panel . ]
we use the condition of thermodynamic equilibrium , eq .
( [ eq : tdstability ] ) , between a diagonal fm stripe and its afm environment to explain these features .
the kinetic energy contribution to the stripe s thermodynamic potential is determined by its non - interacting spectrum consisting of @xmath99 bands ( corresponding to the quantization of transverse electron motion within the stripe ) @xmath164 where @xmath165 is the bandwidth of band @xmath166 @xmath167 the resulting density of states in band @xmath168 is then @xmath169 which together with the chemical potential @xmath64 determines the number of electrons @xmath170 per unit length in the band .
since we are interested in relatively low doping levels we consider the lower @xmath171 bands for which @xmath172 and the non - interacting electronic contribution to the total energy is @xmath173 using these terms , the thermodynamic potential in the fm stripes is @xmath174 where the second term is the magnetic energy , taking into account the structure of the boundaries .
@xmath175 remains the same as for an infinite afm region , eq .
( [ eq : omegaafm ] ) .
[ fig : xfmth ] shows @xmath176 evaluated at the fermi energy @xmath64 which solves @xmath177 as a function of @xmath31 _ and _ @xmath99 . in diagonal stripes of width @xmath99 as imposed by thermodynamic equilibrium .
] points of non - analyticity occur whenever the chemical potential increases beyond the bottom of a band , @xmath178 , so that the carriers begin to fill this additional band .
these non - analyticities result in a corrugated landscape for @xmath179 , shown in fig . [
fig : xfmth ] , whereby several values of @xmath99 may correspond to the same @xmath65 .
thus , a small increase in @xmath6 may drive an abrupt change in @xmath99 but leave @xmath65 constant .
this transition is accompanied by a change in the number of partially filled bands within the stripe , as depicted by the black contours in fig .
[ fig : diastr ] . although our monte - carlo results are not sufficient for constructing the phase diagram , they yield convincing evidence that the phases included in the variational hf calculation are indeed the appropriate variational phases to consider
. some examples of ground states obtained by mc simulated annealing are given in fig .
[ fig : montecarlo ] .
note that the moderate cluster size used in the simulation induces finite - size effects apparent , for example , in the imperfections of the stripe configurations .
these results , in addition to results from other simulations done at other parameter values , agree with the general structure of the phase diagram in fig .
[ fig : phsdgm ] .
moreover , the unrestricted nature of the mc method yields also configurations with afm droplets in a fm background ( bottom right of fig .
[ fig : montecarlo ] ) . as mentioned before ,
such states were not considered in the variational hf approach because of the relative difficulty in calculating their hf energy .
nevertheless , there are reasons to believe , as is confirmed by the simulations , that such a phase indeed exists around the transition line between the striped and uniform fm phases , where @xmath180 . , @xmath181 , @xmath182 .
top right : @xmath183 , @xmath184 , @xmath185 .
bottom left : @xmath186 , @xmath181 , @xmath185 .
bottom right : @xmath187 , @xmath188 , @xmath185 .
dark lines outline fm - afm boundaries.,title="fig : " ] , @xmath181 , @xmath182 .
top right : @xmath183 , @xmath184 , @xmath185 .
bottom left : @xmath186 , @xmath181 , @xmath185 .
bottom right : @xmath187 , @xmath188 , @xmath185 .
dark lines outline fm - afm boundaries.,title="fig : " ] + , @xmath181 , @xmath182 .
top right : @xmath183 , @xmath184 , @xmath185 .
bottom left : @xmath186 , @xmath181 , @xmath185 .
bottom right : @xmath187 , @xmath188 , @xmath185 .
dark lines outline fm - afm boundaries.,title="fig : " ] , @xmath181 , @xmath182 .
top right : @xmath183 , @xmath184 , @xmath185 .
bottom left : @xmath186 , @xmath181 , @xmath185 .
bottom right : @xmath187 , @xmath188 , @xmath185 .
dark lines outline fm - afm boundaries.,title="fig : " ] +
the main finding of this paper concerns the geometry of the low temperature phase - separated state in a two - dimensional double - exchange magnet .
we did not invoke any lattice or orbital degrees of freedom but instead concentrated on the effects of the ubiquitous long range coulomb interaction .
we verified that when the relative area occupied by the fm phase , @xmath189 , is sufficiently large , a _ striped arrangement _
( rather than a droplet super - lattice ) is stabilized .
our results also confirm the expectation , based on a previous analysis of the directional dependence of the fm - afm boundary energy@xcite , that diamond - shaped droplets and diagonal stripes are preferred over their square and bond - aligned counterparts .
indications to this effect are also present in other numerical studies of double - exchange models@xcite .
the stability of the stripe phase should not come as a surprise .
in fact , even in the earlier studies , which considered the continuum limit@xcite , it was noted that the energies of the stripe and droplet configurations can be very close , although no parameter window was found where stripes would correspond to the lowest energy configuration . as a result , it is conceivable , as indeed was shown in ref .
, that a stripe phase may be stabilized , even in this limit , once physics due to some sort of additional degrees of freedom is taken into account . stripes also occur naturally in other models , such as @xmath10 or hubbard@xcite , which involve a competition between the afm nature of a parent undoped state and the kinetic energy of the doped charge carriers .
long - range afm interaction was found to favor stripes in the fm ising model@xcite . regarding the case of a pure double - exchange system with coulomb repulsion considered here ,
it has already been argued@xcite that the correct treatment of the boundaries between the fm and afm regions is likely to tilt the balance in favor of a striped arrangement . in the present work , we considered the experimentaly relevant case of nanometer - size fm inclusions ( comprising only a few lattice periods ) . beside mapping the evolution of the geometry of the inhomogeneous system
, we addressed the long - standing question regarding the stability of free magnetic polarons .
as expected , we find polaronic behavior in the region of small carrier concentration @xmath190 and strong coulomb interaction .
away from this regime , individual magnetic polarons coalesce into larger fm areas .
we were able to span the entire intermediate regime between the coventional phase separation ( where the quantized character of the carrier motion becomes unimportant ) and an array of free magnetic polarons ( for which the notion of thermodynamic equilibrium between fm and afm phases becomes irrelevant ) .
we emphasize that the two main physical ingredients underlying our findings , namely , the quantized electronic motion in small fm regions and the directional dependence of the boundary energy , can be viewed as largely model - independent .
therefore , our present conclusions can be expected to stand for any double - exchange model with a long range interaction , including the case when the latter originates from crystal strain fields@xcite .
the bulk of our study was carried out using a variational hf approximation for the energy of various droplet and stripe phases .
it was supplemented by unconstrained hf calculations on moderate size clusters , implemented via monte - carlo simulated annealing .
the hf approximation is expected to gain accuracy whenever the ratio of electrostatic energy to kinetic energy is small . throughout the range of parameters studied by us , this ratio never exceeds 0.15 . moreover , since we deal with the case of fully polarized electronic spins , the spatial part of the many - body wave - function is antisymmetric .
this fact reduces correlation corrections to the hf result which stem from the tendency of any pair of electrons , owing to their mutual repulsion , to be more distant from each other than the hf wave - function would indicate .
we close with a brief discussion of the experimental situation . to the best of our knowledge ,
a conclusive experimental observation of metallic stripes in phase - separated films of cmr materials is yet to be made .
we note , however , that stripe - like charge ordering on the atomic scale ( charge density wave ) was observed in a variety of manganates .
this includes films with different doping levels@xcite , as well as ceramic@xcite and single crystal@xcite samples in the insulating state above @xmath191 .
in addition , it was suggested@xcite that the phase separated state in a three - dimensional system may acquire a filament structure .
nevertheless , we argue that it would be desirable to synthesize manganate films whose phase - separated state clearly exhibits metallic stripes .
in addition to illustrating our theoretical picture , such systems are expected to display unique and potentially useful properties , some of which were not previously observed .
one of these is an anisotropic conductance , whereby the stripes direction determines a low - resistivity axis , which ought to be amenable to reorientation by , _
e.g. _ , applying a voltage . in general , one expects to find history - dependent resistance and memory effects akin to , and probably more pronounced than those observed earlier in phase separated films@xcite , for which not evidence for stripes was reported .
when the sample composition gets close to the one corresponding to a stable striped arrangement , weak perturbations such as external electric or magnetic fields may be sufficient to change the geometry of the fm regions from droplets to stripes , with a drastic change in transport properties in the form of colossal electroresistance due to dielectrophoresis@xcite and large low - temperature magnetoresistance . which manganate system could potentially exhibit a metallic stripe order ?
in general , in order to look for such a state one is interested to explore the parameter space by changing the average carrier concentration @xmath6 , the metallic area fraction @xmath7 , and the strength @xmath34 of the coulomb interaction@xcite .
while @xmath6 is determined by the dopant concentration , the experimentally measurable quantity @xmath7 depends , in our model , on the ratio of the afm coupling @xmath31 to the hopping @xmath15 .
the latter may be affected by , _
e.g. , _ the choice of the rare earth ion .
an example of a system which apparently allows control over the value of @xmath7 is @xmath192 .
three - dimensional crystals of this compound with @xmath6 between 0.25 and 0.5 , are metallic for @xmath193 , with no signatures of phase separation at low temperatures@xcite . at @xmath194 ,
the system is phase - separated@xcite , and exhibits robust insulating behavior , presumably corresponding to well - separated metallic droplets in an insulating matrix . the properties of the phase - separated state change as one decreases the value of @xmath195 , and at @xmath196 it is possible to observe conduction paths formation and switching as a result of an applied current@xcite .
similar behavior is also found in thin films of the same compound , which at least for sufficiently large values of @xmath195 are phase - separated@xcite , as reflected in their peculiar dielectric and transport properties@xcite .
these findings prompt us to suggest looking for signatures of stripes in @xmath192 films by systematically varying @xmath195 , and with it , as indicated above , the relative area @xmath7 of the metallic phase .
we expect stripes to appear around the point where the areas of metallic an insulating phases are equal to each other .
besides @xmath192 , there are other hole - doped manganate systems which may exhibit a stripe geometry of phase separation , see ref . .
in addition , we expect our results to be relevant for some electron - doped manganates@xcite , as well as possibly for eu - based magnetic semiconductors@xcite .
the hf equations for @xmath84 interacting , spin polarized , electrons , may be written in matrix form as an effective eigenvalue equation , which needs to be solved self - consistently@xcite @xmath197 where the effective hamiltonian matrix is given by @xmath198 the indices @xmath199 , @xmath200 and @xmath201 indicate positions on the lattice , @xmath202 is the single - particle part of the hamiltonian , and @xmath203 is the interaction energy of a particle at site @xmath204 and a particle at site @xmath205 .
@xmath206 are the eigenvectors in lattice - site representation , each indexed by label @xmath207 and with @xmath208 as its eigenvalue .
the self - consistent solution yields the hf ground state energy , given by @xmath209 where the summation is over the @xmath84 states with lowest eigenvalues @xmath208 . for the hamiltonian , eq .
( [ eq : ham_simp ] ) , considered in the present study the single - particle term is @xmath210 with an implicit dependence , given by eq .
( [ teff ] ) , of @xmath211 on the configuration of the core spins @xmath212 .
the second term in eq . ( [ hfsinglepart ] ) reflects the interaction between the conduction electrons and a _ continuous _ neutralizing positive background of density @xmath213 , where @xmath214 is the system area , via the coulomb potential @xmath215 . noting that the eigenvectors are normalized to unity , @xmath216 , and that the @xmath217 and @xmath218 terms in eq .
( [ hhf ] ) are equal and opposite , we are led to analyze the following hf hamiltonian @xmath219 .
\label{eq : fullmatrix}\ ] ] we are interested in cases where the core spins configuration is periodic , such that the system can be divided into @xmath220 unit cells , each containing an identical configuration of spins on @xmath221 sites .
let the super - lattice vectors @xmath222 identify the location of the unit cells .
a position @xmath199 on the lattice can then be written as @xmath223 , where @xmath224 is the position within the unit cell @xmath225 , containing @xmath199 .
the spin periodicity implies that @xmath226 between sites @xmath227 and @xmath228 depends only on @xmath21 , @xmath18 , and the super - lattice vector @xmath229 connecting the two unit cells , _
i.e. _ , @xmath230 . as a consequence of bloch s theorem
this means that the energy eigenfunctions , expressed in the @xmath231 representation , take the form @xmath232 , with eigenenergies @xmath233 , where @xmath234 is defined within the first brillouin zone of the reciprocal super - lattice .
the `` band '' index @xmath168 runs from @xmath235 to @xmath221 , and @xmath236 is normalized to unity within a single unit cell . written in the @xmath237 basis ,
the hamiltonian becomes block diagonal , where the matrix elements of the block connecting states with the same @xmath238 are given by @xmath239 \left[\delta_{ij}\sum_{i'}v_{ii'}^{h } \left|\phi_{bi'}\left(\mathbf{k}'\right)\right|^{2 } -v_{ij}\left(\mathbf{k}-\mathbf{k'}\right ) \phi_{bj}^{*}\left(\mathbf{k}'\right ) \phi_{bi}\left(\mathbf{k}'\right)\right ] .
\label{hfhk}\ ] ] here @xmath240 is the fourier transform of the hopping amplitudes @xmath241 between sites @xmath21 and @xmath18 in unit cells separated by a super - lattice vector @xmath242 .
we also introduced @xmath243 where @xmath244 is the area of a unit cell .
the chemical potential , @xmath64 , is defined by @xmath245 $ ] , with @xmath246 denoting the step function .
the hf ground - state energy of the conduction electrons is @xmath247\\ & \times & \theta\left[\mu-\varepsilon_{b}(\mathbf{k})\right ] -\frac{1}{2}unx\int\frac{d\mathbf{r}}{\left|\mathbf{r}\right| } , \label{eq : fullenergy}\end{aligned}\ ] ] where in the last term we have taken the limit @xmath248 .
this diverging contribution is canceled by the self - interaction of the positive background , evaluated in the same limit @xmath249 consequently , the total energy ( not including the contribution of the anti - ferromagnetic interaction between the core spins ) is given by the sum over @xmath168 and @xmath238 in eq .
( [ eq : fullenergy ] ) . in order to evaluate the matrix elements of the hf hamiltonian
, we need a method to calculate the infinite super - lattice sums in eqs .
( [ eq : kinsum])-([eq : hartreesum ] ) .
the first of these is trivial since hopping is allowed only between nearest - neighbor sites within a unit cell or between adjacent ones . in two dimensions this leaves at most five terms to the sum . on the other hand ,
the coulomb interaction is long - ranged , and an infinite number of terms needs to be included in eqs .
( [ eq : focksum ] ) and ( [ eq : hartreesum ] ) .
the hartree interaction matrix ( [ eq : hartreesum ] ) includes two diverging contributions , one coming from the interaction with the average electronic density and the other from the interaction with the positive uniform background .
the two contributions cancel each other . in order to demonstrate this and extract the remaining finite piece we employ ewald summation
( see appendix [ app : ewald ] ) .
the main identity of this method , directly applicable to the evaluation of the first term in eq .
( [ eq : hartreesum ] ) , is @xmath250 as before , @xmath242 are the super - lattice vectors and @xmath251 is the unit cell area . here , @xmath252 are the reciprocal super - lattice vectors , and @xmath253 is an arbitrary constant , chosen to minimize the number of relevant terms in both sums controlled by the complementary error function @xmath254 .
note that the divergence which stems from summing over large @xmath225 vectors in the left hand side of eq .
( [ eq : ewaldquote ] ) is encoded in the @xmath255 term on the right hand side .
this divergence is canceled by the integral over the whole system in eq .
( [ eq : hartreesum ] ) .
this can be readily seen by using eq .
( [ eq : ewaldquote ] ) with @xmath256 , to write it as @xmath257 consequently we find for the hartree matrix @xmath258 where @xmath259 . when the core spins are arranged in fm droplets separated by an afm ordered background
, the conduction electrons can not hop from one unit cell to the other , _ i.e. _ , @xmath260 .
consequently @xmath261 is independent of @xmath262 , see eq .
( [ eq : kinsum ] ) .
under such a condition it is easy to verify that the hf eigenfunctions and eigenenergies are @xmath262-independent as well . to prove this assertion ,
let us assume that it is true and show that it leads to a @xmath238-independent hf hamiltonian , hence closing the argument self - consistently . since the hartree term in the hf hamiltonian , eq .
( [ hfhk ] ) , depends on @xmath234 only through the hf eigenfunctions it obviously fulfills the requirement . to complete the demonstration we note that the same is true for the fock term since it satisfies @xmath263 moreover , eq .
( [ fockdrop ] ) implies that in the case of fm droplets the calculation of the fock term involves only a finite sum ( over the @xmath221 states within each droplet ) .
this is a direct consequence of the vanishing overlap between electronic states in different droplets .
when the core spins are arranged in a striped configuration , hopping is allowed between unit cells along the direction of the stripes . in other words , if we decompose the super - lattice vectors as @xmath264 , where @xmath265 and @xmath266 are primitive vectors along and off the stripe direction , respectively , then @xmath267 . as a result @xmath261 , depends only on the @xmath262 component along the stripes , _
i.e. _ , @xmath268 .
it follows then , using the same reasoning presented above for the droplet case , that the hf eigenfunctions and eigenenergies depend only on @xmath269 , and the fock term takes the form @xmath270\phi_{bi}^{*}(k_a')\phi_{bj}(k_a ' ) \\ & \times & \sum_{n_a } \frac{1-\delta_{ij}\delta_{n_a,0 } } { |{\bf r}_{ij}+n_a{\bf a}|}e^{in_a ( k_a - k_a ' ) a } , \label{fockstripe}\end{aligned}\ ] ] where @xmath271 is the number of unit cells along the stripe , @xmath272 , and @xmath273 is the vector connecting sites @xmath21 and @xmath18 within a unit cell .
in contrast to the hartree term where the interaction decays slowly , the exponential factor in the fock exchange , eq .
( [ fockstripe ] ) , ensures that the series converges relatively fast .
hence , the infinite sum is well approximated by assuming a long but finite stripe . in our calculation
, we used @xmath274 , and verified that larger values change the ground state energy by an insignificant amount . note that the logarithmic divergence of the @xmath275 sum in the case @xmath276 , is integrable , and vanishes upon the summation over @xmath277 .
the development ( based on ref . ) of ewald s summation method begins with defining the function @xmath278 where the vectors @xmath279 correspond to the @xmath220 points of a two - dimensional lattice of area @xmath214 .
@xmath280 is a periodic function of @xmath281 , with the periodicity of the lattice .
therefore , it can be expanded into the following fourier series @xmath282 where @xmath283 are the reciprocal lattice vectors , and @xmath284 here , @xmath244 is the area of a unit cell . using eqs .
( [ fdef])-([appeniden ] ) and the identity @xmath285 we obtain @xmath286 where the integral was split into two at an arbitrary positive value @xmath253 .
finally , calculating the integrals leads to eq .
( [ eq : ewaldquote ] ) , where the divergent piece of the original sum is given by the @xmath255 term ( and when @xmath287 also the @xmath288 term ) in the new representation . the remaining part of the infinite sums over @xmath252 and @xmath225 is rapidly converging at a rate which is optimized by an appropriate choice of @xmath253 . c. martin , a. maignan , m. hervieu , and b. raveau , phys . rev
b * 60 * , 12191 ( 1999 ) ; j.
f. mitchell , c. d. ling , j. e. millburn , d. n. argyriou , a. berger , and m. medarde , j. appl .
phys . * 89 * , 6618 ( 2001 ) .
d. p. arovas , g. gmez - santos , and f. guinea , phys .
b * 59 * , 13569 ( 1999 ) ; j. l. alonso , l. a. fernndez , f. guinea , v. laliena , and v. martn - mayor , phys . rev .
b * 63 * , 064416 ( 2001 ) ; m. yu . kagan , k. i. kugel , a. l. rakhmanov , and d. i. khomskii , fiz .
nizk . temp .
* 27 * , 815 ( 2001 ) [ low temp . phys . * 27 * , 601 ( 2001 ) ] ; n. d. mathur and p. b. littlewood , solid state comm . * 119 * , 271 ( 2001 ) ; d. khomskii and l. khomskii phys .
b * 67 * , 052406 ( 2003 ) ; r. s. fishman , f. popescu , g. alvarez , j. moreno , t. maier , and m. jarrell , new j. phys . * 8 * , 116 ( 2006 ) . s. k. mishra , r. pandit , and s. satpathy , j. phys . : condens .
matter * 11 * , 8561 ( 1999 ) .
m. uehara , s. mori , c. h. chen , and s .- w .
cheong , nature * 399 * , 560 ( 1999 ) .
m. fath , s. freisem , a. a. menovsky , y. tomioka , j. aarts , and j. a. mydosh , science * 285 * , 1540 ( 1999 ) ; a. biswas , m. rajeswari , r. c. srivastava , t. venkatesan , r. l. greene , q. lu , a. l. de lozanne , and a. j. millis , phys . rev .
b * 63 * , 184424 ( 2001 ) . f. x. hu and j. gao , phys
b * 69 * , 212413 ( 2004 ) ; j. gao and f. x. hu , appl .
* 86 * , 092504 ( 2005 ) ; a. masuno , t. terashima , y. shimakawa , and m. takano , appl .
phys . lett . *
85 * , 6194 ( 2004 ) ; i. pallecchi , l. pellegrino , e. bellingeri , a. s. siri , and d. marr , phys . rev .
b * 71 * , 014406 ( 2005 ) ; a. bhattacharya , m. eblen - zayas , n. e. staley , a. l. kobrinskii , and a. m. goldman , phys . rev .
b * 72 * , 132406 ( 2005 ) ; n. takubo and k. miyano , phys . rev .
b * 76 * , 184445 ( 2007 ) .
e. s. vlakhov , k. a. nenkov , t. g. donchev , e. s. mateev , r. a. chakalov , vacuum * 76 * , 249 ( 2004 ) .
a. asamitsu , y. tomioka , h. kuwahara , and y. tokura , nature ( london ) , * 388 * , 50 ( 1997 ) ; y. yuzhelevski , v. markovich , v. dikovsky , e. rozenberg , g. gorodetsky , g. jung , d. a. shulyatev , and ya .
m. mukovskii , phys .
b * 64 * , 224428 ( 2001 ) ; v. markovich , g. jung , y. yuzhelevski , g. gorodetsky , a. szewczyk , m. gutowska , d. a. shulyatev , and ya .
m. mukovskii , phys .
b * 70 * , 064414 ( 2004 ) ; markovich , g. jung , y. yuzhelevsky , g. gorodetsky , y. m. mukovskii , eur .
j. b * 48 * , 41 ( 2005 ) ; g. garbarino , m. monteverde , c. acha , p. levy , m. quintero , t. y. koo , s. w. cheong , physica b 354 , 16 ( 2004 ) ; l. ghivelder and f. parisi , phys .
b * 71 * , 184425 ( 2005 ) ; h. jain , a. k. raychaudhuri , y. m. mukovskii , and d. shulyatev , appl . phys
. lett . * 89 * , 152116 ( 2006 ) ; h. sakai and y. tokura , preprint arxiv:0803.3922 ( 2008 ) .
m. tokunaga , y. tokunaga , and t. tamegai , phys .
lett * 93 * , 037203 ( 2004 ) .
e. l. nagaev , zh .
pisma red .
* 16 * , 558 ( 1972 ) [ jetp lett . * 16 * , 394 ( 1972 ) ] ; v. a. kashin and e. l. nagaev , zh .
. fiz . * 66 * , 2105 ( 1974 ) [ sov .
jetp , * 39 * , 1036 ( 1974 ) ] .
it was argued@xcite , however , that the combined effects of a long - range coulomb interaction , polarizability , and the interaction of electrons with acoustical phonons may stabilize a slab structure in three dimensions .
n. datta , a. messager , and b. nachtergaele , j. stat . phys . *
99 * , 461 ( 2000 ) ; r. lemaski , j. k. freericks , and g. banach , phys .
lett . * 89 * , 196403 ( 2002 ) .
e. eisenberg , r. berkovits , d. a. huse , and b. l. altshuler , phys .
b * 65 * , 134437 ( 2002 )
. e. l. nagaev , zh .
pisma red .
* 6 * , 484 ( 1967 ) [ jetp lett . *
6 * , 18 ( 1967 ) ] ; zh .
fiz . * 54 * , 228 ( 1968 ) [ sov .
jetp * 27 * , 122 ( 1968 ) ] ; j. magn .
magn . mater . *
110 * , 39 ( 1992 ) , and references therein .
s. satpathy , z. s. popovi , and f. r. vukajlovi , phys .
lett . * 76 * , 960 ( 1996 ) ; j. e. medvedeva , v. i. anisimov , o. n. mryasov , and a. j. freeman , j. phys . : condens .
matter * 14 * , 4533 ( 2002 ) .
e. a. giess , r. l. sandstrom , w. j. gallagher , a. gupta , s. l. shinde , r. f. cook , e. i. cooper , e. j. m. osullivan , j. m. roldan , a. p. segmller , and j. angilello , ibm j. res
. develop . *
34 * , 916 ( 1991 ) ; j. krupka , r. g. geyer , m. kuhn , j. h. hinken , ieee trans .
microwave theory tech . * 42 * , 1886 ( 1994 ) .
t. konaka , m. sato , h. asano , and s. kubo , j. supercond . * 4 * , 283 ( 1991 ) ; j. konopka and i. wolff , ieee trans .
microwave theory tech .
* 40 * , 2418 ( 1992 ) .
e. sawaguchi , a. kikuchi , and y. kodera , j. phys .
jpn * 17 * , 1666 ( 1962 ) ; t. sakudo and h. unoki , phys .
* 26 * , 851 ( 1971 ) .
p. w. anderson and h. hasegawa , phys .
100 * , 675 ( 1955 ) .
v. m. pereira , j. m. b. lopes dos santos , and a. h. castro neto , preprint arxiv:0804.3094 ( 2008 ) .
h. meskine , t. saha - dasgupta , and s. satpathy , phys .
* 92 * , 056401 ( 2004 ) ; h. meskine and s. satpathy , j. phys . : condens .
matter * 17 * , 1889 ( 2005 ) .
m. umehara , phys .
b * 54 * , 5523 ( 1996 ) .
s. kumar and p. majumdar , phys .
* 91 * , 246602 ( 2003 ) ; _ ibid .
_ * 92 * , 126602 ( 2004 ) ; s. dong , h. zhu , x. wu , and j .-
liu , appl .
. lett . * 86 * , 022501 ( 2005 ) ; v. b. shenoy , t. gupta , h. r. krishnamurthy , and t. v. ramakrishnan , phys .
lett . * 98 * , 097201 ( 2007 ) .
e. l. nagaev , phys .
b * 64 * , 014401 ( 2001 ) .
e.g. , _ s. r.white and d. j. scalapino , phys
lett . * 80 * , 1272 ( 1998 ) ; _ ibid . _
* 81 * , 3227 ( 1998 ) ; g. hager , g. wellein , e. jeckelmann , and h. fehske , phys .
b , * 71 * , 075108 ( 2005 ) ; r. citro and m. marinaro , eur .
j. b * 20 * , 343 ( 2001 ) ; _ ibid .
_ * 22 * , 343 ( 2001 ) ; m. yu and h. q. lin , int . j. mod
b * 19 * , 299 ( 2005 ) .
u. lw , v. j. emery , k. fabricius , and s. a. kivelson , phys .
lett . * 72 * , 1918 ( 1994 ) .
s. mori , c. h. chen , and s .- w .
cheong , nature * 392 * , 473 ( 1998 ) ; s. cox , j. singleton , r. d. mcdonald , a. migliori , and p. b. littlewood , nature materials * 7 * , 25 ( 2008 ) ; l. sudheendra , v. moshnyaga , e. d. mishina , b. damaschke , t. rasing , and k. samwer , phys .
b * 75 * , 172407 ( 2007 ) .
j. c. loudon , s. cox , a. j. williams , j. p. attfield , p. b. littlewood , p. a. midgley , and n. d. mathur , phys .
94 * , 097202 ( 2005 ) , and references therein .
renner , g. aeppli , h. m. ronnow , mater .
sci . eng .
c * 25 * , 775 ( 2005 ) .
j. tao , d. niebieskikwiat , m. b. salamon , and j. m. zuo , phys .
lett . * 94 * , 147206 ( 2005 ) . s. dong , h. zhu , and j .-
liu , phys . rev . * b76 * , 132409 ( 2007 ) . regarding the value of @xmath34 , see discussion in sec .
[ sec : model ] .
e.g. _ , i. g. deac , j. f. mitchell , and p. schiffer , phys . rev .
b * 63 * , 172408 ( 2001 ) ; d. n. argyriou , u. ruett , c. p. adams , j. w. lynn , and j. f. mitchell , new j. phys . * 6 * , 195 ( 2004 ) , and references therein r. p. rairigh , g. singh - bhalla , s. tongay , t. dhakal , a. biswas , and a. f. hebard , nature physics * 3 * , 551 ( 2007 ) ; g. singh - bhalla , s. selcuk , t. dhakal , a. biswas , and a. hebard , preprint arxiv:0707.4411 ( 2007 ) . one of the candidates is @xmath289 . in crystalline form , the @xmath195-dependence of its properties was studied extensively [ y. tomioka and y. tokura , phys .
b * 66 * , 104416 ( 2002 ) ] .
results obtained on patterned films at @xmath290 [ t. wu and j. f. mitchell , phys .
b * 74 * , 214423 ( 2006 ) ] suggest current - driven formation of conducting stripes . these include both lightly - doped compounds such as ca@xmath291sm@xmath292mno@xmath293 with @xmath294 [ m. respaud , j. m. broto , h. rakoto , j. vanacken , p. wagner , c. martin , a. maignan , and b. raveau , phys . rev .
b * 63 * , 144426 ( 2001 ) ] , and lamno@xmath293 doped with tetra - valent ions [ _ e.g. , _ la@xmath291te@xmath292mno@xmath293 or la@xmath291ce@xmath292mno@xmath293 , with @xmath295 @xmath4-electrons per mn site ; see p. raychaudhuri , c. mitra , p. d. a. mann , and s. wirth , j. appl . phys . * 93 * , 8328 ( 2003 ) ; g. tan , p. duan , g. yang , s. dan , b. cheng , y .
zhou , h. lu , and z. chen , j. phys . :
* 16 * , 1447 ( 2004 ) ] . | we consider a model with competing double - exchange ( ferromagnetic ) and super - exchange ( anti - ferromagnetic ) interactions in the regime where phase separation takes place .
the presence of a long range coulomb interaction frustrates a macroscopic phase separation , and favors microscopically inhomogeneous configurations .
we use the variational hartree - fock approach , in conjunction with monte - carlo simulations to study the geometry of such configurations in a two - dimensional system .
we find that an array of diamond shaped ferromagnetic droplets is the preferred configuration at low electronic densities , while alternating ferromagnetic and anti - ferromagnetic diagonal stripes emerge at higher densities .
these findings are expected to be relevant for thin films of colossal magneto - resistive manganates . |
as known , many key features of continuous - time chaotic dynamics can be treated in terms of two - dimensional or even one - dimensional discrete - time maps ( see e.g. @xcite ) . among them , most simple and perfect ones are the famous bernoulli map , @xmath0 @xmath1mod@xmath2 , and the tent map , @xmath0 @xmath3 , both defined at the interval @xmath4 . almost any trajectory of both maps is chaotic , that is dense in all this interval . in computer , however , any trajectory finishes at zero after @xmath5 steps only , where @xmath6 is number of bits under use .
this contraction of the discretized phase space , from total @xmath7 points to single point @xmath8 , happens because the maps are non - invertible , i.e. relation between @xmath9 and @xmath10 is not an one - to - one correspondence . but a suitable slight distortion can make a discretized map invertible . evidently , any such map is nothing but a permutation of @xmath7 states ( @xmath11 @xmath12/n$ ] ) .
in particular , the invertible discrete versions of the bernoulli map describe the mersenne digital auto - generators of pseudo - random 0 - 1-sequences ( see e.g. @xcite ) . for example
, the mersenne map in fig.1 represents transitions between 16 states of a particular `` toy '' 4-bit mersenne generator ( @xmath13 ) .
the general formula of such maps is @xmath14 where @xmath15 , @xmath16 , and any of @xmath17 , @xmath18 , equals to either zero or unit , @xmath19 ( hence this is class of @xmath20 different maps ) . to see corresponding unitary evolution matrix
, it is necessary to turn the map picture headfirst and insert zeros and ones in place of the empty cells and crosses , respectively .
the map in fig.1 has two immovable points ( @xmath8 and @xmath21 ) and two cycles ( periodic orbits ) with equal lengths @xmath22 . in applications of much greater mersenne maps
( with @xmath23 @xcite ) , the main issue is construction of those having most long cycles , with @xmath24 .
here , we will be interested in modest problem about statistics of cycles of discrete maps , to be concrete , definite class of discrete tent maps .
example of the discrete bernoulli map ( the mersenne map , see body text ) . ]
first consider the variety of generally asymmetric continuous tent maps ( ctm ) described by @xmath25 where @xmath26 .
the inversion of ( [ ctm ] ) reads as @xmath27 with @xmath28 and @xmath29 representing two variants of the inversion .
let us write , with the help of ( [ ictm ] ) , @xmath30 as a function of @xmath31 and then equate @xmath30 to @xmath31 at various possible values @xmath32 @xmath33 .
thus we obtain @xmath34 linear equations which determine @xmath34 points belonging to all possible periodic orbits ( cycles ) of length @xmath35 . that are either irreducible ( indivisible ) orbits or , if @xmath35 has a divider @xmath36 , composite orbits consisting of @xmath37 repetitions of an irreducible orbit with length @xmath36 .
consequently , @xmath38 where @xmath39 is number of different irreducible periodic orbits with length @xmath36 , and the sum is taken over all the dividers of @xmath35 .
in particular , if @xmath35 is a prime number then ( [ expan ] ) yields @xmath40 since @xmath41 ( there are two immovable points ) .
in general , it is not hard to derive from ( [ expan ] ) that @xmath42}-[l/2]\right)\leq ln_l\leq 2^l-2\,\,\label{any}\ ] ] here @xmath43 $ ] @xmath44 if @xmath45 is even , and @xmath43 $ ] @xmath46 if @xmath45 is odd . at any @xmath47 , according to ( [ any ] ) , @xmath48 at relatively small @xmath45 , from ( [ expan ] ) one finds : @xmath49 , @xmath50 , @xmath51 , @xmath52 , @xmath53 , @xmath54 , @xmath55 , @xmath56 , @xmath57 , @xmath58 , ...
next , consider invertible discrete tent maps ( dtm ) representing discrete analogues of the continuous tent maps ( [ ctm ] ) . to be concrete ,
let us introduce the class of discrete maps defined by @xmath59 & , & \,x < a\,\,;\\ y=\left\{\frac { n(n - x)}{n - a}\right\}-1 & , & \,x\geq a\,\ , \end{array } \label{dtm}\ ] ] here , like in ( [ dbm ] ) , @xmath60 and @xmath61 are integers , @xmath62 @xmath63 ; @xmath64 is analogue of @xmath65 in ( [ ctm ] ) , @xmath66 ; @xmath67\,$ ] is an integer closest to @xmath68 from below ( with @xmath67 $ ] @xmath69 if @xmath70 is integer ) , and @xmath71 means closest integer greater than ( or equal to ) @xmath68 ( that is @xmath72 @xmath69 if @xmath70 is integer ) .
two examples of maps defined by formula ( [ dtm ] ) , at @xmath73 , are shown in fig.2 .
let us prove that ( [ dtm ] ) prescribers invertible maps , i.e. establishes an one - to - one correspondence between @xmath74 and @xmath75 .
it is sufficient to prove that the upper r.h.s . and lower r.h.s .
in ( [ dtm ] ) can not produce one and the same number .
indeed , in such a case we would have @xmath76=y=\left\ { n(n - x^{\prime\prime})/(n - a)\right\}-1\,\,,\ ] ] with some @xmath77 and @xmath78 .
this would mean that @xmath79 ( @xmath80 and @xmath81 are some integers ) , therefore , @xmath82 but , obviously , the latter inequality contradicts the preceding equality .
the proof is finished .
it is easy to see also , that absolute value of a deviation of any dtm ( [ dtm ] ) from its ctm prototype ( [ ctm ] ) , with @xmath83 , ( as well as deviation of ( [ dbm ] ) from bernoulli map ) does not exceed @xmath84 lowest bit only ( @xmath85 , in terms of @xmath86 ) . in this sense , any dtm tends to corresponding ctm , when @xmath87 .
however , the inverted map remains strongly discontinuous , which is a payment for its unambiguity .
hence , in essence , the limit of dtm at @xmath87 does not coincide with corresponding ctm .
naturally , there is no simple rule for the cycles ( periodic orbits ) of the dtms .
a number of various irreducible cycles and lengths of these cycles are extremely irregular functions of @xmath64 and @xmath88 . in
contrary to ( [ expan ] ) , now @xmath89 where @xmath90 is a number of different irreducible periodic orbits with length @xmath35 ( possibly , @xmath91 ) .
for example , at @xmath92 and some three next values of @xmath93 this expansion looks as @xmath94 where the second multiplier ( if any ) in each term is number of different cycles of a length represented by the first multiplier .
in other example , for @xmath95 , @xmath96 typically , a dtm have a long cycle whose length is comparable with the total number of points , @xmath88 , i.e. maximal possible length .
this is illustrated by fig.3 relating to @xmath97 .
quite similar pictures take place also at greater @xmath98 .
we see that practically any dtm has a cycle with length @xmath99 .
such long cycles , whose length @xmath45 is comparable with @xmath88 , can be treated as discrete analogues of the chaotic trajectories of ctm .
the left plot in fig.3 demonstrates highly irregular dependence of the maximal cycle length , @xmath100 , on the asymmetry parameter @xmath93 . at the same time
, the right - hand plot there shows that even slight smoothing of the data produces rather regular results .
therefore , it is reasonable to describe the cycles of dtms in statistical language , considering some specific subclasses of dtms instead of individual maps .
let us consider the family of nearly ( asymptotically ) symmetric dtms determined by the conditions @xmath101 since @xmath102 , we obtain the growing statistical ensemble , which is all the more representative one because all the maps defined by ( [ sym ] ) do tend to the same limit @xmath103 @xmath104
. it would be interesting to investigate such the family of asymptotically symmetric discrete tent maps ( asdtm ) , in comparison with the usual symmetric ctm . here
, simplest statistical characteristics of cycles ( periodic orbits ) of the asdtm will be under our attention .
let @xmath105 designates a density of probability distribution of the cycle lengths in this family of maps .
in other words , practically , @xmath106 where @xmath90 is now * _ summary _ * number of periodic orbits of length @xmath45 in all the maps of the asdtm family .
it is useful to introduce also the quantity @xmath107 that is relative ( probability ) measure of points which belong to all periodic orbits with lengths @xmath108 at all the maps . in reality , with the help of an ordinal pc only , it would take rather long time to obtain all the periodic orbits if @xmath109 .
but our computations , performed at @xmath110 , showed that already @xmath111 are satisfactory values , because next @xmath98 s increases do not change the picture qualitatively ( although , of course , providing better numeric accuracy ) . therefore , it is reasonable to prefer calculations at not high @xmath98 ( @xmath112 ) but apply slight smoothing over @xmath45 s values .
concretely , the plot @xmath113 in fig.3 represents the result of averaging of the exact @xmath113 ( defined by ( [ w ] ) ) over the intervals @xmath114\,$ ] , with @xmath115 . such smoothed probability density , @xmath113 ,
is represented by the curved line at right - hand side of fig.3 .
it is easy to find that the best fitting for it is nothing but the inverse proportional law : @xmath116 where factor @xmath117 ensures the normalization , @xmath118 the dependence ( [ ip ] ) is shown by the straight line , and by smooth curve in the inset which demonstrates good quality of this fitting for short cycles too .
the curved line at left side of fig.3 shows an example of the probability ( [ p ] ) .
it is not smoothed , therefore , formed by many steps with very different heights ( like the famous devil s staircase ) .
its closeness to the thinner straight line , which corresponds to @xmath119 , says that the total @xmath120 points are distributed approximately equally between cycles with different lengths .
this is just about what the approximation ( [ ip ] ) says .
if combining ( [ w]),([p]),([ip ] ) and the approximation @xmath119 , we obtain the estimate of mean number of cycles per one map of the family : @xmath121 ( let us recollect that @xmath90 is * _ summary _ * value for all the asdtm ) . at the same time , as the above examples of map expansions into cycles do show , fluctuations in number of cycles from one particular map to another are significantly greater than the mean value ( [ mnc ] ) . according to these examples , as well to the @xmath105 plots , especial contribution to the fluctuations comes from cycles whose lengths are powers of two
. nevertheless , when rising @xmath98 from 12 to 16 a definite decrease of the fluctuations was noticed .
this observation pushes us to the hypothesis that formula ( [ ip ] ) represents a true asymptotics of the cycles length distribution ( [ w ] ) in the limit ( [ sym ] ) .
there is a simple naive explanation of the hypothetical asymptotics ( [ ip ] ) . in above derivation of the estimate ( [ asympt ] ) for the cycles numbers of continuous tent maps ( ctm ) ,
the factor @xmath34 ( with @xmath45 being cycle lengths ) arises from the two - valued property of their inverse maps ( i.e. due to their irreversibility ) .
since the discrete tent maps ( dtm ) under consideration have univalent inversions , in their case this factor must disappear .
thus one deduces the inverse proportional dependence of number of the cycles ( periodic orbits ) on their lengths , i.e. comes to ( [ ip ] ) ( in other words , cycles of different lengths involve approximately equal amounts of points of the discrete phase space ) . in view of such reasonings
, the inverse proportional law can be expected in case of any sufficiently reach family of unitary dtm , not only the asdtm family defined by ( [ sym ] ) ( moreover , in unitary discrete analogues of multi - modal piecewise linear ctm , not unimodular tent maps only ) . in factual finite - precision computer simulations of more or less general ctm , distinguished from the `` pathological '' map @xmath0 @xmath3 ( about it see introduction ) , e.g. asymmetrical ctm ( [ ctm ] ) or the map @xmath0 @xmath122 with e.g. @xmath123 , it can be expected that the law ( [ ip ] ) must manifest itself sooner than ( [ asympt ] ) . indeed ,
if the latter rule was true , then the very short orbits with lengths @xmath124 only ( recollect that @xmath98 is number of bits under operations ) would take all @xmath125 points of the discrete phase space .
in reality , with no doubts , almost any choice of an initial point results in a long orbit ( `` chaotic trajectory '' ) , whose length @xmath45 is comparable with @xmath88 , while it is hard to hit casually into a short orbit .
to resume , we performed computer statistical analysis of periodic orbits of unitary ( reversible ) discrete tent maps , and found that probability distribution of the orbits lengths well obeys the inverse proportional law ( [ ip ] ) . | the discrete unitary ( reversible ) analogues of the continuous ( irreversible ) tent maps are numerically investigated , in particular , the lengths probability distribution of their periodic orbits .
it is found that its density can be well approximated by the inverse proportional law . |
bose - einstein condensation ( bec ) and superfluidity of liquid @xmath0he in random environments including aerogel and vycor glass are active problems in quantum fluid research . in particular , finding out how spatial confinement affects the bose fluid has stimulated both experimental and theoretical studies . below the @xmath3 temperature of 2.17 k ,
liquid @xmath0he enters the superfluid state and behaves as though it has no viscosity .
superfluidity is a macroscopic quantum phenomenon as well as superconductivity , and understanding both have been one of the major goals of quantum statistical physics .
the various observations of superfluidity was successfully explained by the phenomenological two fluid model , which is based on the idea that the system consists of an inviscid superfluid and a viscous normal fluid . on the other hand , the lambda transition had been thought to be caused by bec , which was confirmed by neutron scattering experiments . with a bec ,
a macroscopic number of particles occupies a single particle ground state and is described by a macroscopic wave function .
the inviscid superflow can be described by this wave function .
however the relation between bec and superfluidity is not completely understood .
although superfluidity and bec are closely related to each other , one is not necessary or sufficient for the other .
for example , in a two - dimensional bose system , kosterlitz and thouless proved that superfluidity can exist even without bec , and the superfluidity was actually observed in @xmath0he films .
the bose system in a random environment might be another good example for studying the relation between bec and superfluidity .
this system has received considerable attention because localization effects allow condensed particles to belong to the normal fluid rather than the superfluid and bec can separate from the superfluid .
the phase diagram of this system has been discussed , showing a specific non - superfluid phase .
thus , studies of this system can reveal the relation between bec and superfluidity .
porous glass such as vycor is often used as a random media in experimental studies .
vycor glass is a from 30% to 70% porous glass containing wormholelike pores , the characteristic diameters of which vary from 30 to 100 . by adjusting the pore size and the adsorbed @xmath0he coverage
, we can change the density of @xmath0he and the superfluid transition . by using torsional oscillators , reppy et al .
interesting features at various pore sizes of vycor glass or coverages were observed , particularly the behavior of the superfluid critical temperature and the temperature dependence of the superfluid density .
the superfluid component has two - dimensional behavior when the pore size of vycor glass is large , and becomes three - dimensional as the pore size is reduced .
the superfluid density in such porous glass is smaller than that of bulk @xmath0he and its critical temperature decreases with the coverage . below a certain coverage ,
the superfluid density can no longer exist , even near 0 k. these results show that superfluidity is broken by the random environment .
it is also important to find out how disorder affects the bec .
bec and its elementary excitation in liquid @xmath0he can be observed by neutron scattering .
bulk @xmath0he has the usual excitations such as phonons , maxons and rotons .
they are collective excitations on bec , the excitations except for phonons being absent above the critical temperature .
dimeo et al . and plantevin et al .
used neutron scattering and a torsional oscillator to measure the elementary excitations and the superfluid transition , respectively , of @xmath0he in porous glass .
surprisingly , the dispersion curve in porous glass was the same as that in the bulk , which means that the disorder does not affect the elementary excitations .
furthermore , these elementary excitations were observed even above the superfluid critical temperature .
hence , the bec might persist in the disordered system even above the superfluid critical temperature .
these results have not been understood completely yet ; thus , the exact relationship between bec and superfluidity remains puzzling .
this problem is also interesting from the following theoretical standpoint . in a bose system confined in a random media ,
the long - range - order correlation due to bec can compete with the disorder , so that the bec critical temperature can be reduced .
huang and meng proposed a model of the three - dimensional dilute bose gas in a random potential that assumed a small coverage of @xmath0he in vycor glass . because it is difficult to formulate the random potential for the porous glass , they used a delta - functional impurity potential and analyzed their model using the bogoliubov transformation and taking an ensemble average .
they found that both bec and superfluidity are depressed by the random potential and the superfluidity disappears below a critical density , even at 0 k , which is qualitatively consistent with the observations by reppy et al .
they also predicted a reentrant transition at low densities ; that is , the superfluid phase enters a normal phase again with decreasing temperatures .
however , the random potential of their model does not include the pore size , and thus it is difficult to quantitatively compare to experimental results for a range of pore sizes .
another model is the bose hubbard model with the random potential . by considering the transfer energy , the on - site repulsion , and the random potential , fisher et al
. found that the bose glass phase can exist with the superfluid phase and the mott insulating phase .
the bose glass phase is similar to the anderson insulating phase in metal . in the bose glass phase ,
the condensed particles are localized and thus do not contribute to superfluidity .
thus , the bose glass phase could influence the collective excitations even above the superfluid critical temperature .
however , the theoretical excitation energy for the bose glass phase disagrees with measurements by neutron scattering experiment , so it is not yet clear whether the bose glass phase has actually been detected .
finally , it should be noted that huang and meng s model can not describe the bose glass phase because the ensemble average makes the system uniform .
few theoretical studies of this random system are quantitatively compare to the experiment .
thus , in this work , we improve huang and meng s model by adding the size dependence of the random potential instead of using their the delta - functional potentials . the strength of the random potential can be estimated by comparing calculated and experimental critical coverages below which the superfluid density disappears , even at 0 k. as a result , our model has no free parameters and can be used for quantitative comparisons to experimental data .
this enables us to determine whether or not our picture of the three - dimensional dilute bose gas in random potential is applicable to a real system .
our formulation can not address this question at high temperatures due to the high number of thermally excited quasi particles .
as far as the condensate density is almost independent of temperature at low temperatures , however , our formulation works well , leading to the following results .
( 1 ) the specific heat agrees quantitatively with experimental data at low temperatures .
( 2 ) because of the random potential , the specific heat is not proportional to @xmath2 , as occurs for phonons , but to @xmath1 .
furthermore , by obtaining the condensate density and the superfluid density , we found the following .
( 3 ) when the total density is sufficiently low , bec can persist even when the superfluid density disappears below that critical coverage .
( 4 ) the random potential causes a reentrant transition of the superfluid phase .
finally , we show why decreasing the open pore density of the vycor glass should allow the reentrant phase to be detected experimentally a brief summary of our paper is as follows . in sec .
ii , we describe our model of the dilute bose gas in a random potential and derive the partition function .
section iii tests our model by quantitatively comparing calculated to experimental specific heats . in sec .
iv , the bec density and the superfluid density are obtained and their characteristics are discussed .
section v is discussion and conclusions .
superfluid @xmath0he adsorbed in vycor glass can be modeled by a three - dimensional dilute bose gas in a random external potential .
the grand canonical hamiltonian is @xmath4 \hat{\psi } ( \vec{x } ) \nonumber \\ & & { } + \frac{v_0}{2 } \int dx^3 \hat{\psi}^{\dagger } ( \vec{x } ) \hat{\psi}^{\dagger } ( \vec{x } ) \hat{\psi } ( \vec{x } ) \hat{\psi } ( \vec{x } ) , \end{aligned}\ ] ] where @xmath5 is the field operator for bose particles of mass @xmath6 , @xmath7 is the number operator , @xmath8 is the chemical potential , and @xmath9 is the external random potential that represents the effect of vycor glass . the first term of the hamiltonian is the kinetic energy and the external potential , whereas the second term refers to the hard - sphere interaction between particles with @xmath10 being the coupling constant with the s - wave scattering length @xmath11 .
this repulsive interaction prevents all particles from being localized at the minimum of @xmath12 .
this has similarities to the fermi system with disorder ; for example , fermions can not localize in a single orbital in space due to the pauli exclusion principle .
therefore , the fermion system is stable even if it is free from the repulsive interaction . on the other hand , to prevent the system from collapsing into the minimum of @xmath9 , the bose system should include a repulsive interaction .
this makes the problem more complicated than that of the fermi system .
proceeding in a standard fashion , we introduce the free particle annihilation and creation operators @xmath13 and @xmath14 .
we assume that the level with @xmath15 is macroscopically occupied with occupation number @xmath16 , so @xmath17 and @xmath18 are replaced by a c - number @xmath19 . by making a fourier transformation and neglecting all off - diagonal terms @xmath20 and @xmath21
, we obtain @xmath22 \hat{a}_{\vec{k}}^{\dagger } \hat{a}_{\vec{k } } + \sqrt{\frac{n_0}{v } } \sum_{\vec{k } \neq 0 } \big ( u_{\vec{k } } \hat{a}_{\vec{k}}^{\dagger } + u_{-\vec{k } } \hat{a}_{\vec{k } } \big ) \nonumber \\ & & { } + \frac{1}{2 } v_0 n_0 \sum_{\vec{k } \neq 0 } \big ( \hat{a}_{\vec{k}}^{\dagger } \hat{a}_{-\vec{k}}^{\dagger } + \hat{a}_{\vec{k } } \hat{a}_{-\vec{k } } \big ) , \end{aligned}\ ] ] where @xmath23 is the volume of the system , @xmath24 is the number density of condensate , and @xmath25 is the fourier transformation of @xmath26 . by neglecting the off diagonal terms , we are neglecting the interactions between the excited particle and the random potential and that between pairs of excited particles ; these become important as the temperature rises and the condensate density decreases . hence , this approximation is poor when many particles are thermally excited . nevertheless , this approximation is useful at low temperatures where the condensate density is almost independent of temperature .
all results here are obtained for these low temperatures .
this hamiltonian can be diagonalized by the bogoliubov transformation @xmath27 then the coefficients @xmath28 , @xmath29 and the quasiparticle spectrum @xmath30 are given by @xmath31 next , we take an ensemble - average to quench the random potential .
the random potential simulates vycor glass with a characteristic pore size @xmath32 as follows .
the quenched potential @xmath25 may decay above the characteristic wave number @xmath33 .
thus , we assume the averaged potential @xmath34 , \label{eq : ensemble - av}\ ] ] where av denotes the ensemble - average .
@xmath35 , with dimension @xmath36 , is the characteristic strength of the random potential .
equation ( [ eq : ensemble - av ] ) makes our model completely different from huang and meng s , and we will show that the results are also different .
the coherence length of the bec is thought to be from hundreds to thousands of , whereas the spatial scale of disorder is the pore size in the glass , which is dozens of .
hence , the macroscopic wave function of bec is not sensitive to disorder in and between pores but instead depend on the disorder averaged over the coherence length .
hence , the ensemble - averaged system can become nearly uniform . for a uniform bose system
, it has been proven that the elementary excitation spectrum becomes the gapless goldstone mode .
thus , we set @xmath37 in eq .
( [ eq : gapspectrum ] ) . the resultant diagonalized and ensemble - averaged hamiltonian is @xmath38 @xmath39 , \\
\epsilon_r & = & \frac{m \sqrt{a } n_0^{3/2 } r_0}{4 \sqrt{\pi } \hbar^2 } \big [ - \e^{2 \alpha } ( 5 + 4 \alpha ) \{1 - \erf(\sqrt{2 \alpha}\ } \nonumber \\ & & \hspace{21 mm } + \sqrt{\frac{2}{\pi \alpha } } ( 1 + \alpha ) \big ] , \\
\alpha & = & \frac{4 \pi a n_0}{k_p^2 } , \end{aligned}\ ] ] where @xmath40 is the hard sphere interaction energy at 0 k , similarly , @xmath41 is that for the random potential .
the quasiparticle spectrum @xmath42 is the same as that in the hard sphere bose gas model and is independent of the random potential .
this independence is confirmed by neutron scattering experiments , which justifies the above assumption of @xmath37 ; conversely , if @xmath43 , the spectrum would depend on the random potential .
this hamiltonian enables us to obtain the grand partition function @xmath44 and various physical quantities .
the condensate density is defined by the following relation : @xmath45 where @xmath46 is the particle number density .
the second term represents the noncondensate particle number as @xmath47 @xmath48 , \\
\lambda & = & \sqrt{\frac{2 \pi \beta \hbar^2}{m } } , \quad \theta = \frac{8 \pi a \hbar^2 \beta n_0}{m } , \quad t^2 = \frac{\hbar^2 \beta}{2 m } k^2 .
\end{aligned}\ ] ] [ eq : realcond ] here @xmath49 is the noncondensate density excited by the hard sphere interaction , @xmath50 is the density due to the scattering of condensate particles with the random potential , and @xmath3 is the thermal de broglie wave length . when @xmath11 vanishes , @xmath50 becomes infinite .
this means that the system would collapse if there were no repulsive interactions between particles . because superfluidity is described by the two fluid model , the particle density @xmath46 consists of the normal fluid density @xmath51 and the superfluid density @xmath52 .
the superfluid density @xmath52 can be calculated by linear response theory . because of its viscosity , only the normal fluid responds to a small , applied velocity field .
thus the normal fluid density can be defined by the response of the momentum density @xmath53 to the external velocity field @xmath54 .
linear response theory gives the following relation : @xmath55 \rangle , \\
j_i(\vec{x},t ) & = & \frac{\hbar}{2i } \big\ { \hat{\psi}^{\dagger } ( \vec{x},t ) \frac{\partial \hat{\psi } ( \vec{x},t)}{\partial x_i } \nonumber \\ & & \hspace{7 mm } - \frac{\partial \hat{\psi}^{\dagger } ( \vec{x},t ) } { \partial x_i } \hat{\psi } ( \vec{x},t ) \big\ } , \\ \hat{\psi } ( \vec{x},t ) & = & \e^ { \mathrm{i } ( \hat{h } - \mu \hat{n } ) t / \hbar } \hat{\psi } ( \vec{x } ) \e^ { - \mathrm{i } ( \hat{h } - \mu \hat{n } ) t / \hbar } , \end{aligned}\ ] ] where @xmath56 is the heisenberg field operator .
the static susceptibility @xmath57 is defined as @xmath58 because of the rotational invariance , the static susceptibility @xmath57 can be written @xmath59 where @xmath60 and @xmath61 are the longitudinal and transverse parts , respectively .
the transverse susceptibility @xmath62 is the normal fluid mass density .
the superfluid number density @xmath52 is @xmath63 .
the susceptibility @xmath62 can be calculated by the bogoliubov transformation in eq .
( [ eq : bogoliubov ] ) . after some tedious calculations
, the resultant superfluid density is given by @xmath64 [ eq : super ] where @xmath65 is the normal fluid density due to the elementary excitations , and @xmath66 is that due to scattering with the random potential .
the density @xmath65 can be also obtained using khalatnikov s method that is based on galilean invariance .
the relation @xmath67 shows that the random potential causes the larger normal fluid density than the noncondensate density ; some condensate particles are captured by the random potential to participate in the normal fluid .
this makes it possible to destroy superfluidity even at 0 k when @xmath66 becomes comparable to @xmath46 .
this formulation can be used to obtain various physical quantities including the condensate density , the superfluid density , and the specific heat .
in this section , we compare the calculated specific heat and the superfluid density with experimental results .
quantitative agreement is shown to be good at low temperatures , which supports our assumption of a dilute bose gas in the random potential .
furthermore , we show that the random potential leads to as - yet unobserved behavior of the specific heat . to make a quantitative comparison , we give the following numerical values to the parameters : @xmath68 and @xmath69 are the mass and the s - wave scattering length of a @xmath0he atom .
other parameters are from the experiments of reppy et al .. the volume @xmath23 of open pores in the vycor glass ( about 40% of the total volume of the vycor glass ) is about 1 @xmath70 .
the particle density @xmath46 of @xmath0he inside the vycor glass is estimated as follows . in vycor glass ,
the atoms are adsorbed and fully cover the surfaces of the open pores due to the van der waals attraction .
the pore area is about @xmath71 @xmath72 .
the rest of atoms , which do not participate in the first - layer solid , can behave as a dilute gas inside the pores .
the particle density @xmath46 of the dilute gas is obtained by subtracting the adsorbed amount from the total amount .
this density is estimated to be from 0.001% to 70% of the density of bulk liquid @xmath0he @xmath73 .
because the first layer of @xmath0he adsorbed on the surfaces can not move and behaves as a solid , we assume the pore size 30 of vycor glass is effectively reduced by @xmath74 .
thus , @xmath32 is estimated to be 20 .
the last parameter @xmath35 , which is the strength of the random potential , can be fixed by comparing to experiment . shown in fig .
[ zerosuper](a ) are the data of zero temperature superfluid signals taken in an experiment that used a torsional oscillator ( fig .
12 of reference ) . because the superfluid component does not contribute to the moment of inertia , the resonant frequency and the period of oscillation differ from those without superfluid .
the period difference @xmath75 is approximately proportional to the superfluid component .
is the resonant period difference in a torsional balance experiment , which is approximately proportional to the superfluid density .
, title="fig : " ] + ( a ) is the resonant period difference in a torsional balance experiment , which is approximately proportional to the superfluid density .
, title="fig : " ] + ( b ) here , the superfluid density is nearly proportional to @xmath75 and disappears at a coverage of 17.5 mg .
figure [ zerosuper](b ) shows the superfluid density at 0 k from eq .
( [ eq : super ] ) . as in the experiment ,
the superfluid density becomes zero at a certain coverage that depends on @xmath35 .
thus , the value of @xmath35 can be fixed using the comparison with fig .
[ zerosuper](a ) ; i.e. , @xmath76 . here
, we define @xmath77 , which is the single particle energy converted from @xmath35 . in the vycor glass , @xmath78 is about @xmath79 k. just above the critical coverage , the superfluid density increases linearly for both the experiment and the calculation ; however , their slopes can not be compared because the amplitude of @xmath52 is unknown in the experiment .
because all parameters are now fixed , we will quantitatively compare calculations to experiments .
the specific heat can be obtained from temperature differentiation of the free energy : @xmath80 \ } , \\
\omega_r & = & v ( \epsilon_r + \frac{n}{v } u_0 ) . \end{aligned}\ ] ] [ eq : specific ] where @xmath81 is the free energy of the elementary excitation and the hard sphere interaction , and @xmath82 is the free energy from the random potential .
figure [ speheat ] compares our results to data of low temperature specific heat taken from fig . 1 of reference .
[ speheat](a ) , which shows the data for high density , the density @xmath46 is fixed from the experimental coverage , whereas we fix the density from the superfluid critical temperature in fig .
[ speheat](b ) at low density .
this is because we have no information about the data on the coverage .
figure [ speheat](b ) also shows the superfluid density . ) , @xmath83 and @xmath84 . here
@xmath85 is the ratio of the coverage to the full pore coverage . in ( b ) , calculated and experimental superfluid densities are compared ( @xmath86 ) .
, title="fig : " ] + ( a ) ) , @xmath83 and @xmath84 . here
@xmath85 is the ratio of the coverage to the full pore coverage . in ( b ) , calculated and experimental superfluid densities are compared ( @xmath86 ) .
, title="fig : " ] + ( b ) the theoretical results agree quantitatively with experiment without using free parameters . above 1.0 k in fig .
[ speheat](a ) , the calculated condensate density begins to decrease rapidly ; here , our criterion of constant condensate density fails , which likely causes the discrepancy with experiment
. however , fig .
[ speheat](b ) shows that the calculated specific heat agrees with experiment up to temperatures near the superfluid critical temperature ; in this temperature region , the calculated condensate density hardly decreases .
this means that the system is more dilute than that of fig .
[ speheat](a ) and thus is affected by the random potential rather than the elementary excitations .
we discuss this in the next section .
nevertheless , these comparisons show that our model is accurate at low temperatures .
our model predicts an effect at low temperature that is due to the random potential but has not yet been observed .
this is shown in fig .
[ logspe ] , which is the log - log plot of fig . [ speheat](a ) .
when the system is free from the random potential , the specific heat should increase a @xmath2 because of the contribution from the phonons .
however , with a random potential , the dependence is linear in @xmath1 at low temperatures .
this means that at low temperatures , the contribution from the random potential is larger than that from the elementary excitations ( phonons ) .
the free energy @xmath82 in eq .
( [ eq : specific ] ) depends on the temperature only through the condensate density @xmath87 .
@xmath82 is the energy from the scattering of the condensate particles with the random potential , and the resultant specific heat is given by the energy that the condensate particles need to slip out of the random potential .
experimental observation of this @xmath1-linear dependence might clearly identify the influence from the random potential .
data of the specific heat at @xmath88 .
two lines for @xmath89 and @xmath90 are added . ]
this section describes some characteristic behavior of the condensate density @xmath87 and the superfluid density @xmath52 derived from our model
. figure .
[ 3d ] shows the dependence of @xmath87 and @xmath52 on temperature and density .
both @xmath87 and @xmath52 decrease with decreasing density , even at 0 k. this means that the effect of the random potential on @xmath87 and @xmath52 becomes larger as the density is reduced .
[ 3d ] shows clearly the difference between @xmath52 and @xmath87 . below the critical density
, the superfluid density disappears , although the condensate persists .
( a ) and @xmath52 ( b ) at @xmath91 . , title="fig : " ] + ( a ) ( a ) and @xmath52 ( b ) at @xmath91 . , title="fig : " ] + ( b ) this situation indicates that the condensate particles can not move as a superfluid because they are trapped by the random potential . we expect that this theoretical result will be confirmed by measurements of condensate density . figure .
[ reent ] shows the temperature dependence of the superfluid density just before superfluidity disappears .
this figure shows the reentrant transition at which the superfluid density @xmath52 goes to zero with a decrease of temperature . in this temperature region , the condensate density is almost constant ; hence , our formulation should work well in accordance with the criterion described in sec .
this reentrant transition is understood as follows . at low temperature and low density . in this regime
, the superfluid density @xmath52 goes to zero with a decrease of temperature ( reentrant transition ) . ]
the condensate depletion @xmath50 of eq .
( [ eq : realcond]c ) comes from the scattering of the condensate particles with the random potential .
this decreases with @xmath87 as the temperature rises , so that @xmath92 of eq . (
[ eq : super]c ) also decreases . as shown in fig .
[ reentreason ] , the magnitude of this decrease in @xmath66 exceeds the increase in @xmath65 of eq .
( [ eq : super]b ) , which is the normal fluid density due to the elementary excitations , in the very low temperature region @xmath93 mk . in other words , condensate particles that are trapped by the random potential at lower temperatures can escape at higher temperatures and thus participate in superfluidity . and
@xmath66 near the reentrant distribution at @xmath94 ( dashed line in fig .
[ reent ] ) . ]
this reentrant transition has not been observed experimentally , probably because it should only occur at very low densities and low temperatures .
however , large values of @xmath35 can make the reentrant transition observable as follows .
we define the temperature @xmath95 as that which maximizes the superfluid density @xmath52 , and define @xmath96 .
[ reentreason2 ] shows the density dependence of @xmath95 and @xmath97 .
both variables increase with @xmath35 . and
@xmath97 for two values of @xmath35 .
@xmath95 is the temperature that maximizes the superfluid density @xmath52 , and @xmath98 .
, title="fig : " ] + ( a ) and @xmath97 for two values of @xmath35 .
@xmath95 is the temperature that maximizes the superfluid density @xmath52 , and @xmath98 .
, title="fig : " ] + ( b ) therefore , the reentrant transition is more likely to be measured at larger @xmath35 .
the parameter @xmath35 is the strength of the random potential over the entire space , and one way to increase @xmath35 is to decrease the open pore density of the vycor glass .
the present paper describes the dilute bose gas system in a random potential .
the outcomes of our studies are as follows . by including the pore size dependence of vycor glass in the random potential
, our model could closely match the experimental conditions of liquid @xmath0he in vycor glass .
we fixed the strength of the random potential by equating the theoretical and experimental critical coverages below which the superfluid density at 0 k vanishes . no other parameters could be adjusted , and thus we could quantitatively compare theory to experiment for other physical quantities .
first , we showed that the calculated specific heat for vycor glass quantitatively agrees with measurements .
this indicates that liquid @xmath0he in vycor glass behaves as a dilute bose gas in a random potential . for low temperatures ,
the calculated specific heat was linear in @xmath1 because of the random potential .
second , the bec was shown to persist even when superfluidity disappears below the critical density . finally , we showed that the reentrant transition of the superfluid phase is more likely to be observed experimentally by increasing the strength of the random potential . because we neglected interactions between pairs of excited particles and between excited particles and the random potential
, this model does not apply to systems at high temperatures . to overcome this limitation
, we are improving the model to include these interactions and will report on this more general model in the near future .
99 d.r.tilley and j.tilley , _ superfluidity and superconductivity _ ( institute of physics publishing bristol and philadelphia , 1990 ) , third edition .
p.e.sokol , in _ bose - einstein condensation _ ( cambridge university press , 1995 ) , edited by a.griffin , d.w.snoke and s.stringari , p.51 .
k.huang , _ statistical mechanics _
( john wiley and sons , 1987 ) , second edition .
j.m.kosterlitz and d.j.thouless , j.phys .
* c6*(1973)1181 .
d.j.bishop and j.d.reppy , phys .
. lett . * 40*(1978)1727 n.trivedi , in _
computer simulation studies in condensed matter physics v _ , ( springer , berlin , 1993 ) ; w.krauth , n.trivedi and d.ceperley , phys .
* 67*(1991)2307 .
k.huang , in _ bose - einstein condensation _ ( cambridge university press , 1995 ) , edited by a.griffin , d.w.snoke and s.stringari , p.31 .
j.d.reppy , j. low temp .
phys . * 87*(1992)205 .
b.c.crooker , b.hebral , e.n.smith , y.takano and j.d.reppy , phys .
lett . * 51*(1983)666 .
r.m.dimeo , p.e.sokol , c.r.anderson , w.g.stirling , k.h.andersen and m.a.adams , phys .
lett . * 81*(1998)5860 .
o.plantevin , b.fk , h.r.glyde , n.mulders , j.bossy , g.coddens and h.schober , phys . rev .
* b63*(2001)224508 .
c.j.pethick and h.smith , _ bose - einstein condensation in dilute gases _ , ( camdridge university press , 2002 ) . k.huang and h.f.meng , phys .
lett . * 69*(1992)644 .
m.p.a.fisher , p.b.weichman , g.grinstein and d.s.fisher , phys . rev . *
p.w.anderson , phys . rev . *
w.krauth , n.trivedi and d.ceperley , phys .
j.s.langer and t.neal , phys .
lett . * 16*(1966)984 .
l.p.gorkov , a.i.larkin and d.e.khmelnitskii , jetp lett .
* 30*(1979)248 s.stringari , in _ bose - einstein condensation _ ( cambridge university press , 1995 ) , edited by a.griffin , d.w.snoke and s.stringari , p.86 .
p.c.hohenberg and p.c.martin , ann .
( n.y . ) * 34*(1965)291 .
i.m.khalatnikov , _ introduction to the theory of superfluidity _ , ( benjamin , new york , 1965 ) , p.13 . | there is a growing interest in the relation between bose - einstein condensation ( bec ) and superfluidity .
a bose system confined in random media such as porous glass is suitable for studying this relation because bec and superfluidity can be suppressed and controlled in such a disordered environment .
however , it is not clear how this relation is affected by disorder and there are few theoretical studies that can be quantitatively tested by experiment . in this work ,
we develop the dilute bose gas model with a random potential that takes into account the pore size dependence of porous glass .
then we compare our model with the measured low - temperature specific heat , condensate density , and the superfluid density of @xmath0he in vycor glass .
this comparison uses no free parameters .
we predict phenomena at low temperatures that have not yet been observed .
first , the random potential causes a @xmath1-linear specific heat instead of the @xmath2 dependence that is usually caused by phonons .
second , the bec can remain even when the superfluidity disappears at low densities .
and third , the system makes a reentrant transition at low densities ; that is , the superfluid phase changes to the normal phase again as the temperature is reduced .
this reentrant transition is more likely to be observed when the strength of the random potential is increased . |
the understanding of polymer crystallization and its theoretical description in the framework of polymer statistics remains a challenging problem since its discovery @xcite-@xcite .
the classical theories of polymer crystallization @xcite-@xcite ( see also @xcite-@xcite and citations therein ) are based on the nucleation theories developed for low molecular weight systems .
the lamellar thickness is identified ad hoc with the critical size of nuclei , and is not related with the key features of polymers in melts and solutions . despite enormous interest in polymer crystallization over many decades , the molecular mechanism of polymer crystallization is not understood , and a polymer statistics related description is not available .
a review of the research in the field of polymer crystallization in the eighties can be found in @xcite .
many important specific features of polymer crystallization were established in the recent research in the field , which includes experimental work @xcite-@xcite , computer simulations @xcite-@xcite , and theoretical studies @xcite-@xcite ( and citations therein ) .
the progress in the field of polymer crystallization in recent years is reviewed in @xcite-@xcite .
a new view of polymer crystallization was developed by strobl @xcite-@xcite , who started from the observation that the nucleation based theories are in disagreement with experiments @xcite - rastogi91 .
the dependence of the crystallization temperature @xmath2 on the inverse lamella thickness @xmath3 has a larger slope than that of the melting temperature , so that these curves intersect at some temperature ( comparable with the temperature of zero growth @xmath4 ) , which implies that the polymer crystallization can develop only below @xmath4 . the picture proposed by strobl
is based on a multistage character of the crystallization process , and the existence of a mesomorphic layer as precursor to the crystalline phase without however specifying the statistical mechanical origin of his scenario .
the aim of this article is to develop a description of polymer crystallization on a more microscopic level relying on polymer specific properties , e.g. , the coil structure of the polymers , and in terms of the relevant microscopic interactions between the monomers in the supercooled polymer melt .
i develop a kinetic mechanism of chain folding in polymer crystallization which is based on the competition between the formation of stems and the random coil shape of polymer chains , and is based on the view that the driving force is due to the repulsions between fluctuational stems , which form below a characteristic temperature @xmath5 due to favored occupation of the trans states along chains , and orient in order to minimize the excluded volume .
the importance of repulsive interactions in this approach is in accordance with the general evidence of the role of repulsive interactions in liquid - solid phase transformations .
the dynamic interplay between forming stems and rouse dynamics considered in the present approach is expected to manifest itself as specific fluctuations prior formation of polymer crystals , and might be responsible for the mesomorphic layer postulated by strobl .
the relevance of density and orientation fluctuations for polymer crystallization is not new and is the start point of the spinodal decomposition theories for the description of early stages of polymer crystallization in @xcite , @xcite . to cite @xcite `` these ( _ density and orientation _ )
fluctuations are caused by an increase in the average length of rigid trans segments along the polymer backbone during the induction period '' .
however , the lamellae thickness in these approaches was not brought in connection with polymer properties .
the article is organized as follows .
section [ lamella ] introduces the basic ideas of the kinetic mechanism of chain folding .
section [ gr ] introduces to the calculation of the growth rate .
section [ concl ] summarizes our conclusions .
fluctuational occupations of the trans states in an supercooled polymer melt of interpenetrating chains result in formation of stems possessing a finite lifetime .
the repulsion between the neighbor stems forces them to orient parallel to each other in order to minimize the excluded volume and results in formation of bundles .
the increase of average occupation of trans states below a temperature @xmath5 increases the lifetime of stems , and enforces the effect of repulsions .
the orientations of stems due to their mutual repulsions is similar to the mechanism of the isotropic - nematic transition in lyotropic liquid crystals , where according to onsager @xcite the minimization of the excluded volume is responsible for the transition .
the difference between liquid crystals and polymers is that in the case of polymer crystallization the stems do not exist from the beginning , but emerge due to occupation of trans states , and orient and grow due to repulsive interactions between the stems .
the present approach is in accordance with simulations in @xcite-@xcite , where chain folding takes place in polymer crystallization by taking into account only repulsive intermolecular interactions .
the intramolecular dihedral energies are associated with the differences between , e.g. , gauche and trans states , and are responsible for the formation of stems .
the length scale of forming fluctuational stems , which are forced to orient due to mutual repulsive interactions , is determined by the competition between the growth of stems and their relaxation due to rouse dynamics . for the average formation time of stems
we adopt the following phenomenological expression @xmath6 where @xmath7 is the supercooling and @xmath8 and @xmath9 are constants .
the term in ( [ fs1 ] ) which is proportional to @xmath10 means that nonzero supercooling is necessary for stem growth .
the non - zero value of @xmath8 accounts for the effect of an orienting crystal surface on the formation of stems and bundles , and is thus legitimate for secondary crystallization and in heterogeneous nucleation at small supercooling , where the crystal growth begins at seeds .
( [ fs1 ] ) yields for the ratio @xmath11 the expression @xmath12 for the longitudinal growth rate of stems .
a similar expression for the lateral growth rate for small supercooling is well - known in the literature ( see for example @xcite ) .
because stems at the time scale @xmath13 are expected to form and decompose , the quantity @xmath14 , which has the dimensionality of velocity rate , is a fluctuational quantity , so that the above estimate has to be understood as the typical value of @xmath14 .
the rouse time of a polymer with arc length @xmath1 is given by @xmath15 where @xmath16 is the monomer friction coefficient ( @xmath17 for polyethylene ) , and we have replaced @xmath2 in ( [ trouse ] ) by @xmath5 ( @xmath18 for polyethylene ) , which is legitimate for small supercooling .
the quantity @xmath19 is the monomer diffusion coefficient .
the balance between the stem growth and their spatial orientations , which is determined by the coil structure of the polymer , can be expressed as follows @xmath20 resolving ( [ ts ] ) with respect to @xmath1 we obtain the characteristic length scale determined by the interplay between the stem growth and the random coil structure of polymer chains as @xmath21 where @xmath22 .
( [ dl - ineq ] ) gives an estimate of the lower bound of the lamellar thickness .
the experimental lamellar thickness is always larger than that given by eq .
( dl - ineq ) , because two subsequent stems along a polymer can fold or form one stem with some probabilities ( see fig .
[ coil - and - stems ] ) .
other processes e.g. those responsible for lamellar thickening also result in increase of @xmath1 . resolving eq .
( [ dl - ineq ] ) with respect to @xmath2 we arrive at the following relation between the crystallization temperature and the lamellar thickness @xmath23 the @xmath24 relation without replacing @xmath2 in the expression of the rouse time by @xmath5 reads @xmath25 .
let us compare the orientation of stems in polymer crystallization with the isotropic - nematic transition in lyotropic liquid crystals , where the transition is determined by the condition @xmath26 , where @xmath27 and @xmath28 ( @xmath29 is the density of rods , @xmath30 their length , and @xmath31 the transverse size ) are the decrease of orientation entropy and the increase of the translational entropy , respectively . because the forming stems in the polymer melt overlap , the above onsager condition is fulfilled and the stems can directly orient due to repulsive interactions .
the van der waals interactions will be enhanced in bundles and contribute to their stabilization .
the present kinetic mechanism of folding favors the switchboard fold surface , which is the consequence of the coil structure of polymers .
the formation of stems and lamellae can be visually interpreted as space segregation of the trans and gauche states .
it is intuitively expected that the repulsive forces promote a local disentanglement of interpenetrated chains , while the attractive van der waals interactions favor an amorphous state , and are therefore expected to be only important for stabilization of the bundles and consequently of the lamellar structure .
the van der waals forces between polymer pieces outside the bundles and with the fold surface would result in their adsorption on the latter . therefore , we expect that the ( logarithmic ) thickening of bundles and lamellae is also caused by repulsive interactions .
the enthalpic contributions to the processes of polymer crystallization are due to i ) the energy difference between trans - and gauche states , and ii ) the difference of the contribution of van der waals forces in the amorphous and the lamellar crystalline states .
the condition @xmath32 is a consequence of the assumption @xmath33 in the ansatz in eq .
( [ fs1 ] ) .
the van der waals attraction of formed stems to the surface is expected also to contribute to the non - zero value of @xmath8 .
the estimate of @xmath34 for @xmath35 nm yields for example for polyethylene the value of order of @xmath36 . setting @xmath13 given by eq .
( [ fs1 ] ) equal to @xmath34 expresses the interplay between the isotropic - nematic ordering of emerging stems and the coil structure of polymers .
the qualitative picture is that that the stems grow until the parts of a chain outside a bundle , which are also forced to form stems , will be located on the lateral side of the bundle , and will likely belong to the same bundle ( see figure [ coil - and - stems ] ) .
thus , the competition between the growth and orientation of stems and the coil structure of the polymer chains yields that the consecutive stems along the polymer fold and belong to the same bundle .
therefore , the longitudinal growth of the bundles is restricted by the coil structure of polymers , which applies in the rouse theory on all scales .
the slower ( logarithmic ) thickening of bundles can further occur . at larger supercooling
, a smaller stem length will be selected as a result of this interplay .
the kinetic mechanism of chain folding suggests that the lamellar thickness @xmath1 follows the change of supercooling .
the latter is in accordance with observations made long ago @xcite .
the chains in polymer melts are gaussian coils irrespective of the presence of entanglements , so that according to the above picture we expect that entanglements weakly influence the condition ( [ ts ] ) .
however , the influence of entanglements on the growth rate is more complex ( see @xcite , vol ii , sect . 9.14 and @xcite ) .
the proposed folding mechanism is expected to apply for both primary and secondary crystallization processes . for primary crystallization , where the surface effect is marginal ( @xmath37 ) , one can apply ( [ t - dl1 ] ) with @xmath38 .
we expect that the above folding mechanism applies for crystallization from polymer solution too , where the slow collapse due to van der waals interactions occurs first , which is followed by the fast folding mechanism due to repulsion - orientation coupling after the repulsive interactions become significant as a result of increase of density .
we also expect that the above folding mechanism based on trans and gauche conformation states is generic for polymers with more complicated local conformation states .
( [ ts ] ) implies that the folding length @xmath1 is selected in the fluctuational regime associated with the ( ` microscopic ' ) time @xmath39 . because the ansatz given by eq .
( [ ts ] ) is local , the lamellar thickness is expected , in contrast to the growth rate , to be robust with respect to changes of external parameters such as molecular weight , etc .
this consequence of the ansatz in eq .
( [ ts ] ) is in accordance with experiments @xcite .
the lamella thickness @xmath1 lies for melts or solutions in the range between @xmath40 nm , and shows a weak dependence on the moderate increase of pressure wunderlich63 , whereas the melting temperature significantly increases under pressure .
the influence of external pressure on polymer crystallization shows wunderlich64 , @xcite , vol.ii , fig .
12.8 ( spherulites ) that the melting temperature increases with pressure .
the lamellar thickness increases smoothly for moderate pressures @xcite . at large pressures
the lamellar thickness increases considerably and can achieve a few microns , and approaches that of an extended crystal @xcite .
a hexagonal phase was observed for large pressure @xcite , which has at the coexistence curve a lower density than the orthorhombic phase @xcite , p. 171 .
these observations emphasize the importance of the repulsive interactions in polymer crystallization .
this is similar to the van der waals gas where the increase of the pressure shifts the interplay between the repulsive and attractive interactions in favor of the former .
the responsibility of the repulsive interactions for stability of the hexagonal phase in polymer crystals is similar to the formation of the triangular lattice of flux lines in type ii superconductors , which has its origin in repulsive interactions between the flux lines @xcite .
the repulsive interactions in the spatially ordered stems in the lamellae facilitates the sliding diffusion , which is expected to be responsible for the increase of lamellae thickness at large pressures @xcite .
thus , the increase of the lamellar thickness , the development of the hexagonal phase at high pressures are direct evidences of the importance of repulsive interactions for stem formation in polymer crystallization .
since @xmath41 and because of fact that crystallization can occur for @xmath42 , the crystallization line , which is described by eq .
( t - dl1 ) , has to cross the melting line , which is given by the gibbs - thomson relation @xmath43 where @xmath44 is the heat of fusion and @xmath45 is the surface tension of the fold surface .
note that eq .
( [ t - dl1 ] ) with @xmath32 is in accordance with strobl s analysis of experimental data .
the parameters in eq .
( [ t - dl1 ] ) can be estimated from the fit to the crystallization line using the experimental data from @xcite , which is shown in figure [ t - dl - fit ] .
-caprolactone ) from @xcite . dashes : our fit of the crystallization line . filled circles and the continuous line : the melting line . ]
the ratio @xmath46 is equal to @xmath47 and possesses for poly(@xmath48-caprolactone ) according to fig .
[ t - dl - fit ] the value @xmath49 k. the slope of ( [ t - dl1 ] ) , which is obtained as @xmath50 , possesses according to [ t - dl - fit ] the value @xmath51 k nm .
thus , eq . ( [ t - dl1 ] ) with constants @xmath52 and @xmath53 estimated for poly(@xmath48-caprolactone ) as @xmath54 and @xmath55 coincides with the crystallization line in fig .
[ t - dl - fit ] ( fig . 10 in @xcite , and fig .
11 in heck - strobl - njp99 ) .
it is likely that uncertainties of the experimental data are responsible for not too excellent fit in fig .
[ t - dl - fit ] .
we now will consider the time evolution of formation and deposition of stems in the vicinity of the crystallization front .
a self - consistent treatment of the mutual correlations of different stems enables one to consider a time - dependent growth rate of one stem @xmath56 . at the time @xmath57 , when , on average , one stem attaches to the crystal surface , the quantity @xmath58 is the growth velocity given by the ratio of one attached stem per time @xmath59 .
the multistage character of the growth means that the growth occurs by sequences of processes , and implies that @xmath60 depends on @xmath56 , which enables one to write down the following phenomenological differential master - type equation for @xmath56 as function of time @xmath61 where @xmath62 ( @xmath63 in the following ) and @xmath64 is a constant .
the factor @xmath65 in ( [ g1 ] ) takes into account the slow - down of the variation of @xmath66 with time , which is expected to be due to the random character of the attachment process .
the absence of a positive term on the right - hand side of ( [ g1 ] ) ensures that eq .
( [ g1 ] ) does not possess a steady state solution .
note that master - type equations for crystal growth ( see e.g. @xcite ) describe many stems growth , and do possess a steady state solution .
( [ g1 ] ) is similar to the multiplicative renormalization of quantities from microscopic to macroscopic scales in the theory of critical phenomena in the case when the coupling constant does not renormalize @xcite .
because eq . ( [ g1 ] ) is intended to describe the growth rate of one stem it should be integrated from @xmath67 until the time @xmath57 , which corresponds to formation and attachment time of one stem to the crystalline front .
the integration of eq .
( [ g1 ] ) from @xmath67 until the time @xmath57 yields @xmath68 for comparison , the growth rate for ballistic deposition is independent of @xmath69 , while @xmath66 for diffusion controlled deposition is proportional to @xmath70 .
the above equation suggests that @xmath59 is infinite at the onset of crystallization , where the growth rate is zero i.e. @xmath71 at @xmath72 .
note that in contrast to ( [ g1a ] ) with @xmath73 the expression for the longitudinal growth rate , which follows from eq .
( [ fs1 ] ) , is given by @xmath74 .
a naive identification of @xmath59 with the time given by the condition ( ts ) yields @xmath75 however , as consequence of the intersection of the crystallization and melting lines given by eqs .
( [ t - dl1]-[gt ] ) shown in fig .
t - dl - fit the growth rate is non - zero only below the intersection temperature @xmath76 .
the latter is also in accordance with strobl s analysis of experimental data @xcite , where the growth rate becomes zero at the temperature @xmath77 .
thus , to take this circumstance into account we adopt ( [ ts1 ] ) with @xmath78 replaced by @xmath79 which yields the time @xmath80 , which is larger than @xmath59 .
the difference between @xmath59 and @xmath80 can be understood as follows : while @xmath81 ( @xmath82 ) gives the selection rule for lamella thickness from the comparison of time scales of competing processes ( i ) stem formation and ( ii ) coil shape of polymers , and does not make a statement on the time course of the growth process , the time @xmath80 is associated with the real time of formation and attachment of a stem at the crystal surface by taking into account the complicated dynamics , and is therefore much larger than @xmath83 . note that the orientation time of stems is not included in @xmath59 , but in @xmath84 . inserting @xmath85 for @xmath59 in ( [ g1a ] ) we arrive at the following estimate of the growth rate @xmath86 where @xmath87 is introduced to take into account the increase of the relaxation time ( viscosity ) with decrease of temperature .
it follows from eq .
( [ g3 ] ) that @xmath66 possesses a pronounced maximum as a function of @xmath88 .
note that in contrast to the turnbull - fisher expression @xcite for @xmath89 in nucleation theory , the above expression describes the attachment rate of one stem .
to obtain the experimentally measured growth rate one should multiply ( [ g3 ] ) with the average number of stems formed per time and per volume .
the quantity @xmath90 is equal to the characteristic temperature @xmath91 appearing in the growth rate @xmath92 given by eq .
( 6 ) in @xcite .
the fit for poly(@xmath48-caprolactone ) yields @xmath93 , so that one obtains @xmath94 .
note that the existence of two separated time scales @xmath39 and @xmath85 , which have a clear physical meaning in the present approach , is in accordance with the experimental finding that the growth rate depends exponentially on @xmath85 and @xmath64 , which is expected to depend considerably on external parameters such as pressure , molecular weight , entanglements , etc .
@xcite , while the lamellar thickness is determined by the ` microscopic ' time @xmath39 , which , as it follows from the definition , is less sensitive to the external parameters .
to summarize , i developed a kinetic mechanism for polymer folding in polymer crystallization , which is based on the competition between the stem formation , which is the consequence of preferential occupations of trans states in the supercooled polymer melt , their orientation to minimize the excluded volume , and the coil shape of polymer chains .
in contrast to the phenomenological nucleation based theories , where the size of critical nucleus is identified ad hoc with the lamellar thickness , our approach has the aim to understand the basic features of polymer crystallization in terms of microscopic interactions between the monomers in supercooled polymer melts , and the coil structure of polymers .
the present approach suggests that the selection of lamella thickness in polymer crystallization is of kinetic origin , and is determined by the ` microscopic ' time scale @xmath39 .
the ( logarithmic ) lamellar thickening as well as the increase of the lamellar thickness with pressure substantiate the importance of the repulsive interactions in polymer crystallization .
the dynamic interplay between forming stems and rouse dynamics considered in the present approach is expected to manifest itself as an ordered precursor prior formation of polymer crystals , and might be responsible for the mesomorphic layer postulated by strobl .
note that the present kinetic mechanism of chain folding is in accordance with results of very recent molecular dynamics simulation @xcite on the time sequence of basic processes in formation of the crystalline order : `` first the chain segments align , then they straighten , and finally the cluster become denser and local positional and orientational order are established '' .
further , the finding in @xcite that entanglements do not affect the nucleation but the growth process is also in accordance with the main conclusion of the present work that there are well separated time scales responsible for selection of the lamellar thickness and formation and attachment of stems i.e. the crystal growth .
the growth rate for formation and deposition of one stem at the growing surface is derived from the differential master - type equation for the scale dependent attachment rate of one stem , which incorporates the multistage character of the secondary crystallization .
the implementation of the picture of polymer folding in polymer crystallization proposed in this article in coarse grained analytic and numeric models will allow more quantitative predictions on polymer crystallization .
i acknowledge partial financial support from the german research foundation , ste 981/3 - 1 and sfb trr 102 .
i am grateful to w. paul , k. saalwchter , c. schick , and t. thurn - albrecht for useful discussions . | i develop a kinetic mechanism to explain chain folding in polymer crystallization which is based on the competition between the formation of stems , which is due to frequent occupations of trans states along the chains in the supercooled polymer melt , and the random coil structure of the polymer chains .
setting equal the average formation time of stems of length @xmath0 with the rouse time of a piece of polymer of the same arc length @xmath1 yields a lower bound for the thickness of stems and bundles .
the estimated lamellar thickness is inversely proportional to the supercooling .
the present approach emphasizes the importance of repulsive interactions in polymer crystallization , which are expected to be responsible for the logarithmic lamellar thickening and the increase of lamellar thickness with pressure .
an expression for the growth rate for formation and deposition of stems is derived by considering the growth as a dynamic multistage process . |
most of the theoretical models for associative memory of neural networks as typified by the hopfield model @xcite have been based on the idea of rate coding , which assumes that information is coded in the firing rate of a neuron at a particular time . on the other hand
, some experimental results on the visual cortex have been suggesting the possibility of another scheme of information coding , that is , the concept of temporal coding which assumes that information is coded in the relative timing of the firing pulses @xcite . stimulated by these experimental findings , studies of the temporal coding are becoming an active area of theoretical brain researches . recently , there have been an increasing number of papers reporting neurophysiological experimental findings of synchronization phenomena of neurons in animal s brain @xcite .
synchronized firings of neurons can be considered to play a key role in certain types of information processing such as the binding problem @xcite . in order to get insights into basic properties of the scheme of the temporal coding it will be necessary to study neural network models based on neurons that can be expressed by such nonlinear oscillators as limit - cycle type and integrate - and - fire type oscillators .
indeed , a network of integrate - and - fire neurons , where one can be directly concerned with the spike timing of neurons , is one of such models and was previously studied @xcite to confirm its functioning as associative memory accompanied by synchronized firings
. the simplest theoretical model of synchronization of coupled oscillators will be the case of a system of phase oscillators of kuramoto et al .
@xcite , which is derived under the assumption of weak coupling limit for general types of limit - cycle oscillators . in an associative memory model based on such a system with the hebb type learning rule
, the distribution of native frequencies will become crucial for the behavior of synchronization or desynchronization of the system @xcite .
if all of the oscillators have an identical native frequency and the couplings are symmetric , the network can be reduced , by an appropriate coordinate transformation , to a system that has a lyapunov function ensuring stability of its equilibrium state(s ) , which implies that the network eventually settles into a synchronized state . on the other hand , in the case when native frequencies are distributed , lyapunov functions in general no longer exist except for a particular case @xcite and the network becomes essentially of dynamic nature ; one may expect partial synchronization or more complex oscillatory behavior . in the present paper ,
we report results of our detailed analysis of associative memory of an oscillator neural network model with distributed native frequencies .
whereas oscillator networks with a lyapunov function were studied previously by several authors @xcite , satisfactory analysis of the case without a lyapunov function has been far less conducted @xcite .
the oscillator network model we consider is described by the following phase evolution equations @xcite : @xmath1 where @xmath2 denotes the total number of neurons , and @xmath3 and @xmath4 respectively denote the phase and the native frequency of @xmath5-th neuron .
we assume the strength of interactions to be given according to the hebb learning rule , @xmath6 where @xmath7 ( @xmath8 ) represents the @xmath9-th stored patterns .
we consider the case when the number of patterns is extensive , @xmath10 . in order to elucidate the effect of the distribution of native frequencies on the behavior of synchronization we want to make the model as simple as possible , and
then we assume @xmath11 to take values @xmath12 rather than continuous values with @xmath13 , ( @xmath14 $ ] ) , which would be more appropriate for the study of temporal coding itself .
furthermore , we deal with a simple case where the native frequency distribution @xmath15 is discrete and symmetric with respect to the central frequency @xmath16 : @xmath17 , @xmath18 . it is noted that without loss of generality one can set @xmath19 as a result of rotational symmetry of the phase evolution equation ( [ phase - eq ] ) .
if the native frequencies are the same , i.e. , @xmath20 independent of @xmath5 , all of the oscillators can get synchronized with @xmath21 . in the case of distributed native frequencies ,
there remain , in general , a group of desynchronized oscillators .
then the total system can not settle into entirely equilibrium states .
assuming that influences exerted by a group of desynchronized neurons on the macroscopic behavior of the total system can be neglected , we use the self - consistent signal - to - noise analysis ( scsna ) @xcite to analytically obtain the macroscopic order parameter equations for the time - stationary states of synchronized oscillations of the network accompanying memory retrieval . introducing the complex variable @xmath22 to express the state of @xmath23-th neuron
, we formally obtain the fixed point equations by putting @xmath24 ; @xmath25 where @xmath26 is the local field and @xmath27 denotes its complex conjugate .
defining the overlap @xmath28 one can rewrite the local field as @xmath29 . within the frame work of the scsna , considering the retrieval solutions of eq .
( [ equilibrium ] ) for which @xmath30 and @xmath31 , we assume the local field to be described as @xcite @xmath32 here the first term involving @xmath33 is the signal part , while the remaining terms represent noise part involving complex gaussian noise @xmath34 ( @xmath35 , real ) together with the effective self - coupling terms proportional to @xmath36 and its complex conjugate @xmath37 .
we note that eq .
( [ equilibrium ] ) yield no solutions if @xmath38 , which does not ensure @xmath39 .
this means that neurons with @xmath38 can not take part in synchronized motions exhibited by neurons with @xmath40 .
although the desynchronized neurons , each of which is expected to oscillate with a certain modified frequency , will make the local fields time dependent quantities , their effect can be expected to cancel out to good approximation provided taking the time average is considered . in further analysis
, we discard the contribution from the desynchronized neurons by setting @xmath41 for @xmath38 .
following the standard procedure of the scsna , we obtain the self - consistent equations for macroscopic variables in the limit @xmath42 .
making use of the rotational symmetry of the phase evolution equation eq .
( [ phase - eq ] ) to choose a gauge such that overlap @xmath33 becomes real i.e. @xmath43_{\{\omega,~\xi^1\ } } = 0 $ ] , we have @xmath44_{\{\omega,~\xi^1 \ } } , \\ & & q = \frac{1}{n}\sum_i ( { \rm re}[x_i])^2= \left[\int { \rm d}u { \rm d}v ( \cos \phi)^2 \right]_{\{\omega , \xi^1 \ } } , \\ & & u_1 = \frac{1}{n}\sum_i \frac{\partial x_i}{\partial z_i } = \left [ \int \frac{{\rm d}u{\rm d}v } { 2 \left\ { ( \xi^1 m + u)\cos\phi+v\sin\phi \right\}+4\gamma_2\cos2\phi}\right]_{\{\omega,~\xi^1 \ } } , \\ & & u_2= \frac{1}{n}\sum_i \frac{\partial x_i } { \partial{\tilde z_i } } = \left [ \int \frac{-\cos2\phi { \rm d}u{\rm d}v } { 2 \left\ { ( \xi^1 m + u)\cos\phi+v \sin\phi \right\ } + 4\gamma_2\cos2\phi } \right]_{\{\omega,~\xi^1 \ } } , \\ & & \gamma_1=\frac{\alpha ( 1-u_1 ) } { ( 1-u_1)^2-u_2 ^ 2 } , \\ & & \gamma_2=\frac{\alpha u_2 } { ( 1-u_1)^2-u_2 ^ 2 } , \\
\label{gammp } & & q_1 = \frac{\alpha q } { ( 1-u_1-u_2)^2 } , \\ & & q_2 = \frac{\alpha ( 1-q ) } { ( 1-u_1+u_2)^2 } , \\ & & { \rm d}u{\rm d}v = \frac{{\rm d}u{\rm d}v } { 2\pi \sqrt{q_1q_2}}\exp \left[-\frac{1}{2 } \left ( \frac{u^2}{q_1}+\frac{v^2}{q_2 }
\right ) \right],\end{aligned}\ ] ] with @xmath45 being implicitly determined by @xmath46 in the above equations , @xmath47_{\{\omega,~\xi^1 \}}$ ] means taking the average over the distribution @xmath15 and the pattern @xmath48 , and the gaussian integration @xmath49 is to be performed over the noise @xmath50 satisfying the condition @xmath40 .
it is noted that in performing the gaussian integration in eqs .
( 6a)-(6i ) , one has to take into account the maxwell rule to pick up the relevant solution @xmath3 , when eq .
( [ cond2 ] ) admits multi - solutions owing to the presence of the self - coupling term ( @xmath51 term ) .
unlike @xmath51 term , @xmath52 term has no contribution to the equilibrium fixed - point equation ( 7 ) . in what follows , for the sake of simplicity we deal mainly with the case
@xmath53 a characteristic feature of the distribution is that the presence of oscillators with central frequency @xmath16 is allowed with a finite fraction @xmath54 and the effect of desynchronized part can be described in terms of @xmath55 and @xmath54 alone .
setting @xmath56 recovers the case that allows an energy function that is bounded from below , @xmath57\equiv -\frac{1}{2}\sum_{ij}j_{ij}\cos(\phi_i-\phi_j)$ ] .
then all of the oscillators get synchronized for large times with the equilibrium configuration @xmath58 determined by eq .
( [ equilibrium ] ) or eq .
( [ cond2 ] ) together with the maxwell rule , which is explained in fig . 1 . analysis based on the scsna of such networks as having lyapunov functions is on the same level of approximations as the replica symmetric calculations @xcite .
figure 2 represents the dependence of @xmath33 on the loading rate @xmath59 computed from the scsna equations ( 6 ) and ( [ cond2 ] ) .
we see that the storage capacity is given by @xmath60 with @xmath61 .
the present value of the storage capacity slightly differs from the result reported previously @xcite , which was obtained by an inappropriate treatment of the maxwell rule @xcite .
we note that the maxwell rule becomes necessary only for @xmath59 in the neighborhood of @xmath62 . under the assumption of random patterns with @xmath63 , the magnitude of @xmath51 plays a crucial role in determining the @xmath62 , as can be seen in fig .
this should be compared with the case of @xmath64 ( @xmath65 $ ] ) , where the order parameter equations involve no @xmath51 term as in the case of cook s model @xcite ( see also recent work in ref .
we note here that when discrete patterns with @xmath66-state variables @xmath67 are considered , the @xmath51 term does not appear in the order parameter equations , if @xmath68 .
the effect of distribution of native frequencies on the behavior of memory retrieval is summarized in fig . 3 , where overlap @xmath33 from the scsna equations is plotted as a function of @xmath59 and @xmath55 for @xmath69 .
there appear two distinct retrieval regimes separated by a valley or gap located around a region with an intermediate value of @xmath55 in the @xmath33-@xmath59-@xmath55 space . in the regime with small @xmath55 , most of the oscillators undergo synchronized motions .
the storage capacities @xmath62 corresponding to the edge of the @xmath33 surface are observed to decrease , as @xmath55 is increased , to attain a certain minimum value , where a crossover to the other regime occurs . in the regime with large @xmath55 ,
oscillators with @xmath55 oscillate with their own frequencies that are modified from the original @xmath55 as a result of entrainment phenomenon .
such a behavior can easily be understood by considering the system in the limit @xmath70 . in the large @xmath55 limit , while most of the oscillators with @xmath71 get synchronized , the oscillators with @xmath72 can not get synchronized and they oscillate with their own frequency @xmath73 independently of the synchronized neurons with @xmath71 .
it is clearly seen in this case that the desynchronized neurons does not contribute to the time - averaged local fields acting on the synchronized neurons .
then the system can be viewed as a diluted system with only a fraction @xmath54 of neurons participating in memory retrieval .
it is interesting to note that the storage capacity @xmath62 in the large @xmath55 region increases as @xmath55 is increased .
the crossover between the two regimes or the gap in the @xmath74 curve with fixed @xmath59 can be more clearly seen in fig . 4 , which display the @xmath55 dependence of overlap @xmath33 obtained from the scsna together with the result of numerical simulations ( @xmath75 ) for a fixed value of @xmath59 .
as the fraction @xmath54 of the oscillators with the central frequency @xmath16 is varied , there occur three types of behavior of the @xmath74 curve showing the cross over between the two regimes . in fig . 5
, we show the @xmath55 dependence of the overlap @xmath33 obtained for @xmath76 and @xmath77 with @xmath78 .
while for @xmath54 large one sees a continuous crossover between the small @xmath55 regime and the large @xmath55 regime , for @xmath54 small the large @xmath55 regime disappears . for only intermediate values of @xmath54
, one observes the gap in the @xmath74 curve to appear .
the occurrence of such two retrieval regimes is not restricted to the simplest case of @xmath79 , but is also observed for more general case with @xmath80 . when a sufficiently large number of oscillators have native frequencies near the central one @xmath16 , the system can behave the same way as that of @xmath79 with @xmath81 , as is shown fig . 6 , where the @xmath82 dependence of the overlap @xmath33 obtained by numerical simulations is displayed for the native frequency distribution with @xmath83 and @xmath84 : @xmath85
indeed we see a gap separating the small @xmath82 regime from the large @xmath82 one .
although there are no oscillators with @xmath19 because of @xmath84 , most of the oscillators get synchronized for the small @xmath82 regime .
we have studied the behavior of the oscillator neural network system with distributed native frequencies . to date analysis of associative memory models
has been mostly confined to the case of networks with an energy or lyapunov function that allows use of replica symmetric calculations of statistical mechanics . in the present work , making use of the method of the scsna that is free from the energy concept @xcite , we have succeeded in analyzing properties of _ the associative memory accompanied by synchronized oscillations _ in the prototype oscillator network that has no lyapunov functions except for the case when @xmath86 .
_ we have shown that the oscillator network can work as associative memory based on temporal coding of simple type even in the presence of distribution of native frequencies . _
our approach has taken advantage of the fact that such temporal attractors as limit - cycles of certain types of dynamical systems can be reduced to a fixed - point type attractors as a result of the system s symmetry property .
the distribution of native frequencies does not allow the present coupled oscillator system to settle into an entirely synchronized state but into a partially synchronized one .
the contributions from desynchronized neurons to the macroscopic behavior of the system have , however , been found to be almost negligible .
thereby , the partially synchronized state of the system has turned out to be almost determined by the long time behavior of the group of synchronized neurons , which can be described by fixed - point type attractors giving rise to retrieval states . in other words ,
_ memory retrieval is achieved by synchronization of oscillatory motions of neurons . _ under the assumed type of native frequency distribution we have found that _ the partial synchronization is classified into a high degree of synchronization that occurs for small @xmath55 with overlap @xmath33 large and a low degree of synchronization that occurs for large @xmath55 with @xmath33 small .
_ finally we note that oscillator neural networks are considered to have advantage over fixed - point type neural networks in several respects .
first , oscillator neural networks exhibit the ability to easily and efficiently discriminate a successful retrieval from unsuccessful one , because the settling into a retrieval state of oscillator neural networks implies the appearance of oscillations with an appreciable amplitude and the central frequency @xmath16 in the overlap , and hence , the local fields of neurons .
second , by utilizing phase as well as amplitude as dynamical variables representing output of a neuron @xcite it becomes possible for neural information to be processed in terms of spatio - temporal patterns of neuronal firings .
in particular , information on time domain is available by employing the scheme of temporal coding @xcite , where assuming uniformly distributed random numbers @xmath87 on @xmath88 for components of the memory patterns , the phase difference @xmath89 between the two oscillators @xmath5 and @xmath90 eventually settles into the difference @xmath91 of the memory pattern @xmath9 .
our present model setting based on the use of binary patterns with @xmath92 @xcite presents a special as well as simple case of the above temporal coding , where the pattern of synchronization is either of in phase or out of phase .
such a case may be related to the problem of segmentation of an object from its background @xcite .
more generally , pattern segmentation @xcite seems to be one of the unique features of oscillator neural networks as has been studied by wang et al .
@xcite , who pointed out that use of limit cycles as attractor states of associative memory facilitates switching between one pattern and another on the time domain .
while studying the important issue of the functionality of pattern segmentation provided by oscillator neural networks will require taking into account some specific ingredients such as the sparseness of the memory patterns and appropriate inhibitory couplings , our simple model will have wide applicability in exploring the computational ability or relevance exhibited by oscillator neural networks from the viewpoint of analytical studies . extending the assumed symmetric native frequency distribution to more general cases of asymmetric one as well as continuous ones
. results of such issues together with details of the present work will be reported elsewhere . *
maxwell rule used in the scsna for picking up the relevant solution for output @xmath36 as a function of @xmath93 and @xmath94 among multi - solutions to eq .
( [ cond2 ] ) . in a graphical representation of solving eq .
( [ cond2 ] ) for which three solutions ( given by the intersections of the two curves ) appear as @xmath93 or @xmath94 is varied , the relevant solution is chosen as the external intersection that delimits the larger area enclosed between the two curves .
such solution is marked by the filled circle .
the maxwell rule ensures the condition of the free energy minimum @xcite in a system with a lyapunov function as in the case of the liquid - vapor phase transition . *
@xmath59 dependence of the overlap @xmath33 obtained from the scsna ( solid line ) together with that from numerical simulations with @xmath95 in the case @xmath96 . to observe the result that the contribution of the @xmath51 term in eq .
( [ local ] ) or ( [ cond2 ] ) to the value of @xmath33 is significant , we plot values of @xmath33 obtained by deliberately setting @xmath97 ( broken line ) . *
phase diagram showing @xmath33-surface plotted as a function of @xmath59 and @xmath55 for the network with distributed native frequencies with @xmath98 .
we observe a valley or gap separating the small @xmath55 regime from the large @xmath55 one .
the dependence of the storage capacities @xmath62 on @xmath55 is represented by the projected curve on the @xmath59-@xmath55 plane .
* dependence of @xmath33 on @xmath55 for the network with @xmath78 and @xmath69 obtained from the scsna ( line ) and numerical simulations with @xmath75 ( dots ) .
the gap separating the two regimes with different types of synchronization are clearly depicted and is in excellent agreement with the results of numerical simulations .
* dependence of @xmath33 on @xmath55 for @xmath78 and @xmath99 obtained from the scsna . * the numerical simulation result ( @xmath95 ) for @xmath83 showing the @xmath82 dependence of @xmath33 in the case with @xmath100 and @xmath78 ( see eq .
( [ l=2 ] ) ) .
r. ritz , w. gerstner and j.l .
van hemmen , `` associative binding and segregation in a network of spiking neurons '' , _ models of neural networks ii _ , eds .
e. domany , j.l .
van hemmen and k. schulten , ( springer - verlag , 1994 ) chap . | we study associative memory of an oscillator neural network with distributed native frequencies .
the model is based on the use of the hebb learning rule with random patterns ( @xmath0 ) , and the distribution function of native frequencies is assumed to be symmetric with respect to its average .
although the system with an extensive number of stored patterns is not allowed to get entirely synchronized , long time behaviors of the macroscopic order parameters describing partial synchronization phenomena can be obtained by discarding the contribution from the desynchronized part of the system .
the oscillator network is shown to work as associative memory accompanied by synchronized oscillations . a phase diagram representing properties of memory retrieval
is presented in terms of the parameters characterizing the native frequency distribution .
our analytical calculations based on the self - consistent signal - to - noise analysis are shown to be in excellent agreement with numerical simulations , confirming the validity of our theoretical treatment . |
in a companion paper , which we will refer to as paper i ( `` the haar wavelet transform of a dendrogram '' ) , a new transform is applied to a hierarchical clustering .
various examples are given of uses of this transform , prior to applying the inverse transform . in this paper
, we look at linkages with other ways of understanding the wavelet transform , with the classical dmarche described in appendix 1 . our aim is to understand the wavelet transform when applied to hierarchical clustering dendrograms ( where notation and expression as ultrametric topology are summarized in appendix 2 ) .
after all , both the wavelet transform and hiearchical clustering aspire to multiresolution or multiscale analysis .
the natural question is then : how do they differ and are there different aspects that they bring to the data analysis task ?
a good deal of recent work on wavelet transform has been through group theoretic approaches .
foote et al . ( 2000a ) point to how group theoretic understanding can lead to a `` wealth of new analysis filters '' ( in the context of multiresolution signal and image analysis ) .
the same point is made by the smart project ( smart , 2005 ) , including the change to have automatic generation of new transform algorithms .
believing that algorithms should be developed if and only if the there is a verifiable user need for them , we would instead point to another reason why group theoretic understanding is crucial to data analysis
. a great deal of observed reality can be understood by way of observed symmetries , and groups summarized and encapsulate the properties of these symmetries . for time evolving phenomena , therefore , or spatial coordinate referenced phenomena , it may be possible to replace analysis that is time - referenced or referenced to particular coordinate systems with a more general , more generic , symmetry analysis .
this is the vision opened up by the study of group actions on a set of objects .
parenthically , one fascinating way as to how this works can be seen in cendra and marsden ( 2003 ) .
the authors develop ( i ) an analytic theory of dynamics as functions of spatial and temporal coordinates ; and ( ii ) group theoretic interpretation of , in parts of the study , return or phase maps .
our approach can be stated as follows .
let @xmath2 be an ultrametric space , associated with an @xmath3-dimensional embedding , @xmath4 .
we note that an ultrametric space is necessarily of 0 dimenensionality ; and that minimal dimensionality real embedding of an ultrametric has been studied by lemin and others ( lemin , 2001 ; bartal et al . , 2004 ) . the partial order of ( clopen ) set inclusion
is denoted by the binary tree or hierarchy , @xmath5 .
consider the group action comprising rotations or cyclic permutations ( these are equivalent ) of subnodes of any node in @xmath5 , and we will denote this group as @xmath6
. then we study the wavelet transform of @xmath7 resulting from the actions of group @xmath6 .
having already discussed the new wavelet transform in paper i , we can give one result relating to it in the context of the group of equivalent representations of @xmath5 as follows .
_ theorem : _ for all @xmath8 equivalent representations of @xmath5 ( here : unlabeled graph isomorphisms ) , the dendrogram haar wavelet transform is unique .
the proof follows from the definition of the wavelet coefficents at each level , @xmath9 ; whereas the equivalent representations of @xmath5 are intra level .
it follows from this theorem that we have a unique matrix representation of a dendrogram .
in this section we will review recent work using wavelet transforms on data tables , and show how our work represents a radically new approach to tackling similar objectives .
approximate query processing arises when data must be kept confidential so that only aggregate or macro - level data can be divulged .
approximate query processing also provides a solution to access of information from massive data tables .
one approach to approximate database querying through aggregates is sampling .
however a join operation applied to two uniform random samples results in a non - uniform result , which furthermore is sparse ( chakrabarti , garofalakis , rastogi and shim , 2001 ) .
a second approach is to keep histograms on the coordinates . for a multidimensional feature space , one
is faced with a `` curse of dimensionality '' as the dimensionality grows .
a third approach is wavelet - based , and is of interest to us in this article .
a form of progressive access to the data is sought , such that aggregated data can be obtained first , followed by greater refinement of the data .
the haar wavelet transform is a favored transform for such purposes , given that reconstructed data at a given resolution level is simply a recursively defined mean of data values .
vitter and wang ( 1999 ) consider the combinatorial aspects of data access using a haar wavelet transform , and based on a multi - way data hypercube .
such data , containing scores or frequencies , is often found in the commercial data mining context of olap , on - line analytical processing . as pointed out in chakrabarti et al .
( 2001 ) , one can treat multidimensional feature hypercubes as a type of high dimensional image , taking the given order of feature dimensions as fixed .
as an alternative a uniform `` shift and distribute '' randomization can be used ( chakrabarti et al . , 2001 ) .
there are problems , however , in directly applying a wavelet transform to a data table .
essentially , a relational table ( to use database terminology ; or matrix ) is treated in the same way as a 2-dimensional pixelated image , although the former case is invariant under row and column permutation , whereas the latter case is not ( murtagh , starck and berry , 2000 ) .
therefore there are immediate problems related to non - uniqueness , and data order dependence .
what if , however , one organizes the data such that adjacency has a meaning ?
this implies that similarly - valued objects , and/or similarly - valued features , are close together .
this is what we do , using any hierarchical clustering algorithm ( e.g. , the ward or minimum variance one ) . without loss of generality , as seen in these figures , we assume that a hierarchy is a binary , rooted tree ; and equivalently that the series of agglomerations involve precisely two clusters ( possibly singleton clusters ) at each of the @xmath10 agglomerations where there are @xmath11 observations .
these @xmath11 observations are usually represented by @xmath11 row vectors in our data table . a significant advantage in regard to hierarchical clustering is that partitions of the data can be read off at a succession of levels , and this obviates the need for fixing the number of clusters in advance .
all possible clustering outcomes are considered .
( remark : of course , relative to any one of the commonly used cluster homogeneity criteria , each partition is guaranteed to be sub - optimal at best . )
in this article , we will denote the agglomeration of two clusters , @xmath12 and @xmath13 , as cluster @xmath14 .
so ( left or right subtree ) nodes in the dendrogram are associated with the child ( elder or younger ) subnodes .
we can define the elder cluster as @xmath12 such that @xmath15 , but we will _ not _ be concerned with whether or not elder corresponds to left , and younger to right . for @xmath11 objects or observation vectors ,
another notation that we can use is that the hierarchy @xmath5 is the set of clusters indexed from 1 to @xmath11 : @xmath16 .
we will always assume in this article , for convenience of exposition and with little loss of generality , that for distinct clusters @xmath17 . whenever the distinction between the following become important
, we will clearly distinguish between them : clusters ; nodes ; sets of objects ; sets of indices of objects ; and p - adic number representation of indices of objects .
the haar algorithm , as discussed in paper i , is as follows : 1 .
take each cluster @xmath14 in turn , proceeding in sequence through @xmath18 .
2 . apply the smoothing function , @xmath19 : @xmath20 .
3 . thereby apply the detail function , @xmath21 : @xmath22 .
4 . return to step 1 until all @xmath23 clusters are processed . for details of how the clusters also take terminal nodes ( objects ) into account , see paper i. now , it is clear from construction that perfect reconstruction of the input data ( alternatively expressed , perfect undoing of the foregoing haar algorithm ) is guaranteed , given all of the following : ( i ) all of the detail function values , ( ii ) the final smooth , @xmath24 , ( iii ) the definition of the dendrogram , and ( iii ) a convention of left and right subtree that allows us to traverse down the tree from @xmath14 to both @xmath12 and @xmath13 . in practice
our objectives are to explore the foundations of two distinct approaches .
both seek a haar wavelet basis .
these two approaches are as follows and can express the 2 input data cases considered in section 4.3 ( `` the input data '' ) of paper i. * wavelet transform in an ultrametic topology : induce the haar basis from the hierarchy @xmath5 that expresses the relationships in a set of ultrametrically related points , @xmath25 .
* wavelet transform on embedded subsets : induce the haar basis from the hierarchy @xmath5 defining a set of subsets of @xmath25 .
in the ultrametric case , each point @xmath26 defines an @xmath3-dimensional vector : @xmath27 . for notational convenience
therefore @xmath28 is either the index , or a vector . in the set of subsets case
, each point @xmath26 can be defined as an @xmath11-dimensional index vector .
thus for example the sequentially second point is defined as @xmath29 .
both practical cases above can be expressed as follows : we carry out a wavelet transform in @xmath30 where @xmath31 is the group of alternative representations of a given hierarchy , @xmath5 .
the points @xmath26 are associated either with vectors in @xmath4 or with an orthonormal vector set in @xmath32 .
( note how @xmath32 is @xmath11-dimensional , whereas @xmath4 is @xmath3-dimensional .
the cardinality of @xmath25 is @xmath11 .
the dimensionality of a feature or attribute space is @xmath3 . )
in this section we look at the wavelet transform on discrete fields , and in particular on @xmath33 .
this is realized through cyclic representations of the affine group of @xmath34 or of @xmath35 . in wavelets
, we are seeking a representation of our data which has `` covariant '' properties relative to scale : for example , for tranlation `` covariance '' , the representation of a shifted signal must be a shifted copy of the representation of the signal ( torrsani , 1994 ) . in the group theory perspective , such `` covariance '' properties ( i.e. with respect to the action of a symmetry group ) are the starting point , and the representation is to be derived from them .
the `` covariance '' group `` turns out to be isomorphic ( up to a compact factor ) to the geometric phase space of the representation '' ( torrsani , 1994 , p. 6 ) .
traditionally , the wavelet transform is covariant with respect to a group action applicable to images , signals , time series , etc . ,
viz . the affine group of the real line , which is a continuous group ( antoine et al . , 2000a ) .
thus a first task is to bypass the need for a continuous group .
the group law of the affine group , in generic form @xmath36 , generates translations and dilations .
the action of the @xmath36 group on @xmath37 means : @xmath38 .
we have the following product : @xmath39 here , the identity is : @xmath40 .
the inverse of @xmath41 is : @xmath42 .
this is a non - commutative lie ( and thus continuous ) group .
flornes et al .
( 1994 ) consider a discrete wavelet transform to begin with , specified on the hilbert space @xmath33 ( where @xmath34 are the integers , @xmath35 are integers mod @xmath0 where @xmath0 is prime for reasons explained below ; and @xmath43 implies finite energy from discrete values , or being square integrable ) .
the haar measure is defined on locally compact groups , permitting integration over group actions or members ; and a locally compact separable group is consistent with the square integrable property .
the group at issue here is the cyclic representation of the affine group ; or its finite analog , the affine group mod p ( foote et al . , 2000a ) .
a discussion of square integrable group representations in the context of time - frequency transforms , including the continuous wavelet transform , can be found in torrsani ( 2000 ) ; and torrsani ( 1994 ) discusses the counterexample case of the rotation group , @xmath44 , on the 2-dimensional sphere , which gives rise to a representation which is not square integrable .
the wavelet transform considered by flornes et al .
( 1994 ) has the following operations : @xmath45 the reason why @xmath0 has to be prime is as follows . consider the p - adic number representation of @xmath35 . for the p - adic respresentation of @xmath35 to be a field , i.e. to have an inverse
, @xmath0 must be prime .
the unitary representation of a group @xmath31 is a mapping into a ( complex ) hilbert space .
flornes et al .
( 1994 ) define the following unitary representation , @xmath46 , on the group , mapping into unitary operations on @xmath33 : @xmath47 in terms of the translation and dilation operators , we have : @xmath48 thus far , a purely discrete wavelet transform is at issue .
however if we take our function values @xmath49 defined on @xmath35 as sampled values from a continuous signal , then problems of interpolation arise .
it simply is not good enough to transform our discrete data independently of awareness of the underlying continuum .
note that this issue is at the nub of where data analysis differs from signal processing . in data analysis ,
mostly a data cloud is taken as given ( potentially leading to a combinatorial perspective ) or as a stochastic realization ( leading to a statistical modeling ) . in signal processing ,
the observed data are samples of an underlying topological or other continuous structure .
it is a prime objective in the signal processing context to keep the processing of the observed , sampled data fully and provably consistent with the underlying continuous structures .
a way to address this issue of interpolation is to use b spline filters ( the simplest example of which is the box function used by the haar wavelet transform ) to smooth the data , thereby `` filling the holes '' between gaps in the sampled values , before dilating .
this is termed pseudo - dilation .
relations ( [ eqntd2 ] ) can be re - expressed as follows ( torrsani , 1994 ) : @xmath50 then the affine multiplication law is verified by @xmath51 ( using relation [ eqntd ] ) : @xmath52 when @xmath53 with @xmath0 prime , then @xmath54 always divides @xmath11 .
the affine group on @xmath33 is well - defined ; @xmath55 and @xmath56 define a representation of this group ; and it turns out that the square integrable property holds . in general for @xmath57
the trous algorithm is used incorporating both continuously defined dilation , and discretization of the function .
we could embed each node of a hierarchical clustering , defined as we always do so as binary , rooted tree , in @xmath58 .
lang ( 1998 ) develops a wavelet transform approach ( including the haar wavelet transform and others ) for such a 2-series local field , or cantor dyadic group . taking each cluster @xmath59 or node in the tree individually is not satisfactory from our point of view , and so we look further for a more pleasing way to process a hierarchy .
wavelet transform analysis is the determining of a `` useful '' basis for @xmath60 with the following properties : * induced from a discrete subgroup of @xmath4 , * using translations on this subgroup , and * dilations of the basis functions .
classically ( frazier , 1999 ; debnath and mikusiski , 1999 ; strang and nguyen , 1996 ) the wavelet transform avails of a wavelet function @xmath61 , where the latter is the space of all square integrable functions .
wavelet transforms are bases on @xmath60 , and the discrete lattice subgroup @xmath62 is used to allow discrete groups of dilated translation operators to be induced on @xmath4 .
discrete lattice subgroups are typical of 2d images ( the lattice is a pixelated grid ) or 3d images ( the lattice is the voxelated grid ) or spectra or time series ( the lattice is the set of time steps , or wavelength steps ) . sometimes it is appropriate to consider the construction of wavelet bases on @xmath30 where @xmath31 is some group other than @xmath37 . in foote ,
mirchandani , rockmore , healy and olson ( 2000a , 2000b ; see also foote , 2005 ) this is done for the group defined by a quadtree , in turn derived from a 2d image . to consider the wavelet
transform approach not in a hilbert space but rather in locally - defined and discrete spaces we have to change the specification of a wavelet function in @xmath63 and instead use @xmath30 .
benedetto ( 2004 ) and benedetto and benedetto ( 2004 ) considered in detail the group @xmath31 as a locally compact abelian group .
analogous to the integer grid , @xmath62 , a compact subgroup is used to allow a discrete group of operators to be defined on @xmath30 .
the property of locally compact ( essentially : finite and free of edges ) abelian ( viz .
, commutative ) groups that is most important is the existence of the haar measure ( ward , 1994 ) .
the haar measure allows integration , and definition of a topology on the algebraic structure of the group .
benedetto ( 2004 ) considers the following cases , among others , of wavelet bases constructed via a sub - structure : * wavelet basis on @xmath60 using translation operators defined on the discrete lattice , @xmath62 .
this is the situation discussed above , which holds for image processing , signal processing , most time series analysis ( i.e. , with equal length time steps ) , spectral signal processing , and so on .
as pointed out by foote ( 2005 ) , this framework allows the multiresolution analysis in @xmath60 to be generalized to @xmath64 for minkowski metric @xmath65 other than euclidean @xmath66 . * wavelet basis on @xmath67 , where @xmath68 is the p - adic field , using a discrete set of translation operators .
this case has been studied by kozyref , 2002 , 2004 ; altaisky , 2004 , 2005 .
see also the interesting overview of khrennikov and kozyref ( 2006 ) . * wavelet basis on @xmath67 using translation operators defined on the compact , open subgroup @xmath35 .
( it is interesting to note that @xmath62 is discrete ; and that the quotient @xmath69 is compact .
in contrast to this , @xmath35 is compact ; and the quotient @xmath70 is discrete . ) * discussed is a wavelet basis on @xmath30 , for a group @xmath31 , using translation operators defined on a discrete subgroup , or discrete lattice . * finally
the central theme of benedetto ( 2004 ) is a wavelet basis on @xmath30 where @xmath31 is a locally compact abelian group , using translation operators defined on a compact open subgroup ( or operators that can be used as such on a compact open subgroup ) ; and with definition of an expansive automorphism replacing the traditional use of dilation . a motivation for the work of benedetto ( 2004 ) and benedetto and benedetto ( 2004 ) is laying the groundwork for the wavelet transform on the adelic numbers ( see appendix 3 ) .
in this work we are content to be less ambitious in regard to number systems below we focus on a particular p - adic encoding of dendrograms ; and we are also less ambitious in regard to wavelet functions staying resolutely with the haar wavelet in this work . our motivation is due to our application drivers .
locally compact abelian groups ( lcag ) are a way to take fourier analysis ( hence a particularly important class of harmonic analysis because so versatile ) into more general settings than e.g. the reals ( although the reals also form a non - compact , but locally compact , abelian group ) .
the duals of members of a locally compact abelian group , defined as unitary multiplicative characters , @xmath71 , also form a locally compact abelian group ( knapp , 1996 ) . the duality pairing @xmath31 and @xmath72 allows for an isometry between @xmath30 and @xmath73 ( antoine et al .
, 2000 ) .
fourier analysis is the study of real square integrable functions that are invariant under the group of integer translations ( see foote et al . , 2000a ) , while abstract harmonic analysis is the study of functions on more general topological groups that are invariant under a ( closed ) subgroup .
it is interesting to compare some global properties of our approach relative to the fourier transform approach applied to decision trees in kargupta and park ( 2004 ) .
the fourier transform lends itself well to a frequency spectrum analysis of binary decision vectors , and the latter can be of importance for supervised classification . on the other hand
, our work makes use of binary trees but in the framework of unsupervised classification .
the wavelet transform shares with the fourier transform the property that frequency spectral information is determined from the data ; and the wavelet transform additionally determines spatial or resolution scale information from the data .
we have found the wavelet transform , as described in this article , to be appropriate for the type of input data that we have considered . in general terms , both we in this work , and kargupta and park ( 2004 ) , have as objectives the filtering and compression of data .
we need affine group action for the wavelet transform , and we have seen above ( section [ discrete ] , `` wavelet transform on discrete fields '' ) that @xmath35 affords us this ; but for an arbitrary discrete field , and for an arbitrary locally compact abelian group , it is tricky to find an affine group .
taking further the flornes et al .
( 1994 ) work , antoine et al .
( 2000 ) consider an infinite locally compact abelian group , @xmath74 ; the restriction of @xmath74 to a lattice @xmath75 ; @xmath76 , an abelian semigroup ; and the actions of @xmath76 on @xmath77 .
based on a pseudodilation ( i.e. , the product of a natural dilation by a convolution operator ) the case of a continuous underlying signal is studied , i.e. the relation between the semigroup acting on @xmath77 , and a continuous affine group acting on @xmath63 .
spline functions are again among the wavelet functions used ( among which is the haar wavelet function associated with the b - spline of order 1 ) .
the aspect of greatest interest to us here in the approach of benedetto ( 2004 ) , and benedetto and benedetto ( 2004 ) , is to define wavelets on @xmath30 , with @xmath31 taken as the p - adic rationals , @xmath68 , and with the p - adic integers , @xmath35 , on which we define translation - like operators .
firstly , we are using therefore @xmath67 , i.e. functions defined on the rationals .
secondly , we use a discrete group of operators on @xmath67 which are not in themselves translation operators , but may be used in an analogous way .
the `` trick '' used is that the quotient @xmath78 is discrete , and this will furnish the translation operators .
an expansive automorphism is also needed in this context , i.e. what we use in analogy with dilation . a number of alternatives for the subset of @xmath78 ( more strictly the quotient of the group dual by the annhilator in the dual of the compact open subgroup ) are discussed by benedetto ( 2000a ) . given our application - driven interest ,
we will not pursue them further here .
what we will do , however , is to look at how the group - based approach of benedetto ( 2000a ) , that for the most part assumes infinite sets , can be tailored for our algorithmic hence finite purposes .
we will therefore look at how we can suitably encode any given dendrogram in terms of @xmath68 or indeed , as will be seen , in terms of @xmath35 .
next we will move on to look at how a lattice - proxy is defined on our encoding , and thereby translation operators .
finally , we will look at how an expansive automorphism can be replaced by expansive mapping in the finite and discrete case . in all of this
, we follow the methodology described by benedetto ( 2000a ) ; but we restrict all aspects to the finite , discrete context . for the group actions , with respect to which we will seek invariance
, we consider independent cyclic shifts of the subnodes of a given node ( hence , at each level ) .
equivalently these actions are adjancency preserving permutations of subnodes of a given node ( i.e. , for given @xmath12 , the permutations of @xmath79 . due to the binary tree , or strictly pairwise agglomerations represented by the hierarchy
, the `` adjacency '' property is trivial .
we have therefore cyclic group actions at each node , where the cyclic group is of order 2 .
the symmetries of @xmath5 are given by structured permutations of the terminals .
the terminals will be denoted here by term @xmath5 .
the full group of symmetries is summarized by the following algorithm : 1 . for level @xmath80
down to 1 do : 2 . selected node , @xmath81 node at level @xmath82 .
3 . and permute subnodes of @xmath9 .
subnode @xmath9 is the root of subtree @xmath83 .
we denote @xmath84 simply by @xmath5 . for a subnode @xmath85 undergoing a relocation action in step 3 ,
the internal structure of subtree @xmath86 is not altered .
the algorithm described defines the automorphism group which is a wreath product of the symmetric group .
denote the permutation at level @xmath9 by @xmath87 .
then the automorphism group is given by : @xmath88 where wr denotes the wreath product .
call term @xmath83 the terminals that descend from the node at level @xmath9 .
so these are the terminals of the subtree @xmath83 with its root node at level @xmath9 .
we can alternatively call term @xmath83 the cluster associated with level @xmath9 .
we will now look at shift invariance under the group action .
this amounts to the requirement for a constant function defined on term @xmath89 . a convenient way to do
this is to define such a function on the set term @xmath83 via the root node alone , @xmath9 . by definition
then we have a constant function on the set term @xmath83 .
let us call @xmath90 a space of functions that are constant on term @xmath83 .
possible bases of @xmath90 that were considered in paper i are : 1 . basis vector with @xmath91 components , with 0 values except for value 1 for component @xmath28 .
2 . set ( of cardinality @xmath92 ) of @xmath3-dimensional observation vectors .
the constant function maps @xmath93 where @xmath94 is the space of complex valued functions on the set term @xmath5 .
now we consider the resolution scheme arising from moving from + @xmath95 to @xmath96 . from the hierarchical clustering point of view
it is clear what this represents : simply , an agglomeration of two clusters called term @xmath86 and term @xmath97 , replacing them with a new cluster , term @xmath83 .
let the spaces of constant functions corresponding to the two cluster agglomerands be denoted @xmath98 and @xmath99 .
these two clusters are disjoint initially , which motivates us taking the two spaces as a couple : @xmath100 . in the same way ,
let the space of constant functions corresponding to node @xmath9 be denoted @xmath90 .
the multiresolution scheme uses a space of zero mean denoted @xmath101 with mean defined on the couple of spaces , @xmath100 : @xmath102 in considering spaces of constant functions , @xmath98 and @xmath99 , we know that the support of these spaces are , respectively , term @xmath103 and term @xmath97 .
so if , instead of the space of zero mean denoted @xmath101 where mean is defined on the couple of spaces , @xmath100 , we considered the mean of the combined support , term @xmath104 term @xmath97 , then the result would be quite different .
we would , in fact , have a cluster - weighted mean value .
let us exemplify a case that satisfies all that has been defined in the context of the wreath product invariance that we are targeting .
it also exemplifies the algorithm discussed in depth in paper i. take the constant function on @xmath98 to be @xmath105 .
take the constant function on @xmath99 to be @xmath106 .
then define the constant function on @xmath107 to be @xmath108 .
next define the zero mean function on @xmath101 to be : @xmath109 in the support interval of @xmath98 , i.e. term @xmath86 , and @xmath110 in the support interval of @xmath99 , i.e. term @xmath97 .
evidently @xmath111 .
following on from the previous subsection , a demonstration that the algorithm allows for exact reconstruction of the data the inverse transform is as follows .
the constant function on term @xmath5 corresponding to the root node is @xmath112 .
the two subnodes of the root node , at levels @xmath85 and @xmath113 , are reconstructed from @xmath114 ( and , as we have seen , we can either use @xmath115 or @xmath116 ) .
we next look at the subtrees whose roots are given by the nodes ( just considered ) at levels @xmath85 and @xmath113 , and these subtrees are necessarily disjoint .
all subnodes of these currently selected nodes are reconstructed using the same algorithm .
this procedure is iteratively continued until the terminals have been dealt with .
comparison with traditional clustering criteria is considered next .
it is clear why agglomerative levels are very problematic if used for choosing a good partition in a hierarchical clustering tree : they increase with agglomeration , simply because the cluster centers are getting more and more spread out as the sequence of agglomerations proceeds . directly using these agglomerative levels
has been a way to derive a partition for a very long time .
an early reference is mojena ( 1977 ) . to see how
the wavelet transform used by us leads to a very different outcome , see paper i , where we describe use of the norms of the @xmath115 vectors .
when we consider agglomerative hierarchical clustering algorithms it is clear that ( i ) clusters are often defined in terms of center of gravity , or mean ; and ( ii ) this allows for defining a ( vector ) difference term between a cluster and its immediate sub - clusters .
it is also clear that the haar wavelet algorithm is close to the method known as median or gower s or wpgmc weighted pair group method using centroids : see table , p. 68
, of murtagh ( 1985 ) . we could also cater for other agglomerative criteria , subject to storing the cluster cardinality values , and develop an algorithm that is close to the haar wavelet one .
constant functions on spaces @xmath98 and @xmath99 remain just as before .
the zero mean functions on space @xmath101 would now be generalized to weight mean ( with weights given by cluster cardinalities ) .
viewed in this light , our work has led us to develop a new storage structure for hierarchical clustering trees , which is particularly beneficial for data filtering objectives .
the novelty of our work resides in two areas : ( i ) we have shown the close association between two classes of multiple resolution data analysis approaches , agglomerative hierarchical clustering algorithms and the wavelet transform ; and ( ii ) our motivation is not at all to construct hierarchical clusterings in a new way but rather to illuminate further inherent ultrametric properties of data ( cf .
murtagh , 2004 ) .
the dendrogram wavelet transform has been seen to be a set of applications of a function applied to set members , or cluster members , associated with nodes of the hierarchy @xmath5 .
the non - singleton clusters comprise the set @xmath117 .
the wavelet function is applied in turn to @xmath118 . in this paper
, we are using the haar wavelet function .
but we could well use others ( e.g. , altaisky , 2004 , uses the morlet wavelet ) . with each @xmath59
there is an associated level function , @xmath119 , which induces a total order on @xmath120 .
we will show that the application of the wavelet function to this sequence of clusters is `` dilatary '' or `` expansive '' in two different ways .
we will return to what these two different ways are in a moment .
the hierarchy @xmath121 contains an inceasingly embedded sequence of subsets ( corresponding to increasingly pruning the branches of the tree ; bouchki , 1996 ) .
we have : @xmath122 .
define term as the set of terminal nodes of a hierarchy , and card the cardinality of a set .
then term@xmath123 with card@xmath124 .
card(term@xmath125 ) @xmath126 .
card(term@xmath127 ) @xmath128 .
@xmath129 card(term@xmath130 ) @xmath131 .
each application of the wavelet function is to the minimal ( non - singleton ) cluster ( again see bouchki , 1996 ) in each @xmath132 .
below , we will see how we _ promote _ all @xmath133 to the corresponding @xmath134 by multiplying clusters ( in a particular algebraic representation ) by @xmath135 , which is of norm @xmath0 .
the `` dilatory '' or `` expansive '' character of our sequence of operations , viz . , application of the wavelet function , comes from ( i ) the sequence of embedded subsets of @xmath5 so we still operate on a cluster , but the data on which we work becomes smaller ; or ( ii ) the sequence of levels at which we work is derived from repeatedly taking the product with @xmath135 of norm @xmath0 . in order to introduce this product with @xmath135 of norm
@xmath0 we first describe the p - adic algebraic representation of the hierarchy . in this representation ,
clusters including singletons , have an operator , denoted @xmath136 .
this operator allows clusters to be defined .
next , we have a null element in @xmath120 in this algebraic representation , and a norm of each @xmath59 .
hence for @xmath137 , we can define @xmath138 .
we have : @xmath139 s.t .
finally , @xmath141 we have @xmath142 .
we will introduce now the one - to - one mapping of clusters ( including singletons ) in @xmath5 into a set of p - adically expressed integers ( a fortiori , rationals , @xmath68 ) .
the field of p - adic numbers is the most important example of ultrametric spaces .
addition and multiplication of p - adic integers , @xmath35 , are well - defined .
inverses exist and no zero - divisors exist .
a terminal - to - root traversal in a dendrogram or binary rooted tree is defined as follows .
we use the path @xmath143 , where @xmath144 is a given object specifying a given terminal , and @xmath145 are the embedded classes along this path , specifying nodes in the dendrogram .
the root node is specified by the class @xmath146 comprising all objects .
a terminal - to - root traversal is the shortest path between the given terminal node and the root node , assuming we preclude repeated traversal ( backtrack ) of the same path between any two nodes . by means of terminal - to - root traversals , we define the following p - adic encoding of terminal nodes , and hence objects , in figure [ fig2 ] . @xmath147 if we choose @xmath1 the resulting decimal equivalents could be identical : cf .
contributions based on @xmath148 and @xmath149 . given that the coefficients of the @xmath150 terms ( @xmath151 ) are in the set @xmath152 ( implying for @xmath153 the additional terms : @xmath154 )
, the coding based on @xmath155 is required to avoid ambiguity among decimal equivalents . a few general remarks on this encoding follow .
for the labeled ranked binary trees that we are considering , we require the labels @xmath156 and @xmath157 for the two branches at any node . of course we could interchange these labels , and have these @xmath156 and @xmath157 labels reversed at any node . by doing
so we will have different p - adic codes for the objects , @xmath158 .
the following properties hold : ( i ) _ unique encoding : _ the decimal codes for each @xmath158 ( lexicographically ordered ) are unique for @xmath159 ; and ( ii ) _ reversibility : _ the dendrogram can be uniquely reconstructed from any such set of unique codes . the p - adic encoding defined for any object set above can be expressed as follows for any object @xmath144 associated with a terminal node : @xmath160 in greater detail we have : @xmath161 here @xmath162 is the level or rank ( root : @xmath23 ; terminal : 1 ) , and @xmath28 is an object index . .
branches are labeled @xmath156 and @xmath157 .
clusters are : @xmath163.,width=340 ] in our examples we have used : @xmath164 for a left branch ( in the sense of figure [ fig2 ] ) , @xmath165 for a right branch , and @xmath166 when the node is not on the path from that particular terminal to the root .
a matrix form of this encoding is as follows , where @xmath167 denotes the transpose of the vector .
let @xmath168 be the column vector @xmath169 .
let @xmath170 be the column vector @xmath171 . define a characteristic matrix @xmath172 of the branching codes , @xmath156 and @xmath157 , and an absent or non - existent
branching given by @xmath173 , as a set of values @xmath174 where @xmath26 , the indices of the object set ; and @xmath175 , the indices of the dendrogram levels or nodes ordered increasingly . for figure [ fig2 ]
we therefore have : @xmath176 for given level @xmath162 , @xmath177 , the absolute values @xmath178 give the membership function either by node , @xmath162 , which is therefore read off columnwise ; or by object index , @xmath28 which is therefore read off rowwise .
the matrix form of the p - adic encoding is : @xmath179 here , * x * is the decimal encoding , @xmath172 is the matrix with dendrogram branching codes and * p * is the vector of powers of a fixed integer ( usually , more restrictively , fixed prime ) @xmath0 .
the tree encoding exemplified in figure [ fig2 ] , and defined with coefficients in equations ( [ eqn1 ] ) or ( [ eqn1b ] ) , ( [ eqn2 ] ) or ( [ eqn3 ] ) , with labels @xmath180 and @xmath181 is not commonly used : zero and one labels are more common .
we required the @xmath182 labels , however , to fully cater for the ranked nodes ( i.e. the total order , as opposed to a partial order , on the nodes ) .
we can consider the objects that we are dealing with to have equivalent integer values . to show that , all we must do is work out decimal equivalents of the p - adic expressions used above for @xmath183 . as noted in gouva ( 2003 )
, we have equivalence between : a p - adic number ; a p - adic expansion ; and an element of @xmath35 ( the p - adic integers ) . the coefficients used to specify a p - adic number , gouva ( 2003 )
notes ( p. 69 ) , `` must be taken in a set of representatives of the class modulo @xmath0 .
the numbers between 0 and @xmath184 are only the most obvious choice for these representatives .
there are situations , however , where other choices are expedient . '' as noted already the wavelet basis on @xmath60 is often induced from the discrete subgroup , @xmath62 .
now for a discrete subgroup we use the dendrogram , @xmath5 .
the addition operation on the group @xmath5 will now be explored . in order to define a group structure on the p - adic encoded objects ,
we require an addition operation .
we do not `` carry and add '' in the traditional way because this does not make sense in this context .
instead we define the following `` average and threshold '' operation for any coefficients ( of values of @xmath170 , as used in equations [ eqn1b ] or [ eqn3 ] ) .
we define the following compositions for such coefficients .
@xmath185 examples from the encoding defined above for @xmath183 ( again with reference to figure [ fig2 ] , and equations [ eqn1 ] or [ eqn1b ] , [ eqn2 ] or [ eqn3 ] ) follows .
@xmath186 @xmath187 @xmath188 @xmath189 @xmath190 informally : in the tree , this addition operation only retains non - zero terms for nodes in the tree strictly _ above _ the first ( i.e. lowest level ) cluster within which the two objects find themselves .
this means that if the two objects only find themselves together for the first time in the same cluster that contains all objects then the result of the addition operation is 0 .
let us use our `` average and threshold '' operation , which we are using as a customized addition , to define clusters .
we will do so by example , taking figure [ fig2 ] as our case study .
we will call the clusters , ranked by increasing node level , @xmath118 as used in the caption of figure [ fig2 ] .
@xmath191 @xmath192 @xmath193 @xmath194 @xmath195 @xmath196 @xmath197 the trivial cluster containing all @xmath11 objects , @xmath146 , is of value 0 in this representation
. definition of null element : on the dendrogram @xmath5 , the set @xmath198 is the null element when using our p - adic encoding ( given in definitions ( [ eqn1b ] ) and ( [ eqn3 ] ) ) and addition operation ( [ eqn4 ] ) .
defining p - adic notation for clusters in this way allows us to define norms of clusters ; or to define p - adic distances between clusters ; or indeed to define p - adic distances between clusters and objects ( singletons , terminals ) .
we will look at these in subsection [ sect54 ] below .
for completeness we will provide a definition of p - adic dendrogram multiplication .
take @xmath199 and let @xmath200 .
the product operation is defined on the formal ( laurent ) power series as : @xmath201 with restriction to the term in @xmath202 .
p - adic dendrogram multiplication will be used below in the definition of the expansive operator : this is multiplication by @xmath135 . thus far ,
we have been concerned with an analytic framework .
now we will induce a metric topology on @xmath5 . to find the p - adic distance
, we look for the term @xmath203 in the p - adic codes of the two objects , where @xmath204 is the lowest level such that the absolute values of the coefficients of @xmath203 are equal .
let us look at the set of p - adic codes for @xmath183 above ( figure [ fig2 ] ) , to give some examples of this . for @xmath153 and @xmath205
, we find the term we are looking for to be @xmath206 , and so @xmath207 . for @xmath153 and @xmath208
, we find the term we are looking for to be @xmath209 , and so @xmath210 . for @xmath208 and @xmath211 , we find the term we are looking for to be @xmath212 , and so @xmath213 .
having found the value @xmath204 , the distance is defined as @xmath214 .
see , inter alia , benzcri ( 1979 ) , and gouva ( 2003 ) , for this definition of ultrametric distance .
examples based on figure [ fig2 ] : @xmath215 since @xmath207 .
@xmath216 since @xmath210 .
@xmath217 since @xmath210 .
examples for clusters from figure [ fig2 ] : @xmath218 .
@xmath219 .
we take for a singleton object @xmath220 , and so the norm of an object is always 1 .
we therefore define the p - adic norm , @xmath221 , of an object corresponding to a terminal node in the following way : for any object , @xmath144 , @xmath222 .
the norm of a non - singleton cluster is defined analogously .
it is seen to be strictly smaller .
we have : @xmath223 . for the expansive operator that we use for dilation
, we will consider product with @xmath135 .
the norm associated with this operator is seen to be @xmath224 .
the operator given by multiplication by @xmath135 therefore has norm or modulus @xmath0 .
the p - adic norm , or p - adic valuation , satisfies the following properties ( schikhof , 1984 ) : 1 .
@xmath225 iff @xmath226 2 .
@xmath227 3 .
@xmath228 we also have : @xmath229 with equality only if @xmath12 is a singleton .
consider the set @xmath230 with its p - adic coding considered above .
take @xmath231 .
( non - uniqueness of corresponding decimal codes is not of concern to us now , and taking this value for @xmath0 is without any loss of generality . )
multiplication of @xmath232 by @xmath233 gives : @xmath234 .
each level has decreased by one , and the lowest level has been lost .
subject to the lowest level of the tree being lost , the form of the tree remains the same . by carrying out the multiplication - by-@xmath135 operation on all objects
, it is seen that the effect is to rise in the hierarchy by one level .
let us call product with @xmath135 the operator @xmath235 .
the effect of losing the bottom level of the dendrogram means that either ( i ) each cluster ( possibly singleton ) remains the same ; or ( ii ) two clusters are merged .
therefore the application of @xmath235 to all @xmath12 implies a subset relationship between the set of clusters @xmath236 and the result of applying @xmath235 , @xmath237 .
repeated application of the operator @xmath235 gives @xmath238 , @xmath239 , @xmath240 , @xmath129 .
starting with any singleton , @xmath26 , this gives a path from the terminal to the root node in the tree .
each such path ends with the null element , as a result of the null element definition ( section 3 ) .
therefore the intersection of the paths equals the null element .
benedetto and benedetto ( 2004 ) discuss @xmath235 as an expansive automorphism of @xmath25 , i.e. form - preserving , and locally expansive . some implications of benederro and benedetto s ( 2004 ) expansive automorphism follow .
for any @xmath12 , let us take @xmath241 as a sequence of open subgroups of @xmath25 , with @xmath242 , and @xmath243 .
this is termed an inductive sequence of @xmath25 , and @xmath25 itself is the inductive limit ( reiter and stegeman , 2000 ) .
each path defined by application of the expansive automorphism defines a spherically complete system ( schikhof , 1984 ; gaji , 2001 ) , which is a formalization of well - defined subset embeddedness .
we now return to our starting point , the haar algorithm given in section 1.4 .
we apply the averaging and differencing operations to each cluster in sequence .
but now , after doing this for cluster @xmath12 , we apply the operator @xmath235 , i.e. the @xmath135 product to the p - adic representation of the dendrogram .
this causes us to move up a level .
this is our enhanced concept of dilation , which we apply to the dendrogram , where we keep the same averaging and differencing operations applied to the cluster in sequence .
in our case we are looking for a new basis for @xmath30 where @xmath31 is the set of all equivalent representations of a hierarchy , @xmath5 , on @xmath11 terminals . denoting the level index of @xmath5 as @xmath9 ( so @xmath244 , where @xmath245 are the positive reals ) , and @xmath246 is the level index corresponding to the fine partition of singletons , then this hierarchy will also be denoted as @xmath121 .
let @xmath25 be the set of observations .
let the succession of clusters associated with nodes in @xmath5 be denoted @xmath247 .
we have @xmath23 non - singleton nodes in @xmath5 , associated with the clusters , @xmath12 .
at each node we can interchange left and right subnodes .
hence we have @xmath8 equivalent representations of @xmath5 , or , again , members in the group , @xmath31 , that we are considering .
so we have the group of equivalent dendrogram representations on @xmath121 .
we have a series of subgroups , @xmath248 , for @xmath249 .
symmetries are given by permutations at each level , @xmath9 , of hierarchy @xmath5 .
collecting these furnishes a group of symmetries on the terminal set of any given ( non - terminal ) node in @xmath5 .
the practical application arises through identifying the @xmath11 terminal nodes with ( i ) @xmath3-dimensional vectors , or ( ii ) @xmath11-dimensional hypercube vertices . on the latter sets of vectors we can also consider an associated permutation representation .
parenthetically , we note that the permutation representation is known as the alternating or zig - zag permutations and are counted by the andr or euler numbers ( murtagh , 1984a ; sequence a000111 in sloane , 2005 ) . in this work
we ignore another form of equivalent representation , i.e. that arising from two or more level values being identical : @xmath250 for some @xmath251 .
this means that successive nodes can be interchanged .
this situation happens when we have equilateral triangles in the ultrametric space , as opposed to triangles that are strictly isosceles with small base . at each non - singleton cluster ,
@xmath12 , we define a ( trivial ) affine group on @xmath252 .
the group is defined on @xmath253 matrix interpretation of the haar dendgrogram wavelet transform --------------------------------------------------------------- consider any hierarchical clustering , @xmath5 , represented as a binary rooted tree .
for each cluster @xmath14 with offspring nodes @xmath12 and @xmath13 , we define @xmath254 through application of the low - pass filter @xmath255 : @xmath256 the application of the low - pass filter is carried out in order of increasing node number ( i.e. , from the smallest non - terminal node , through to the root node ) .
next for each cluster @xmath14 with offspring nodes @xmath12 and @xmath13 , we define detail coefficients @xmath257 through application of the band - pass filter @xmath258 : @xmath259 again , increasing order of node number is used for application of this filter .
see paper i for further details .
we now return to the issue of how we start this scheme , i.e. how we define @xmath260 , or the `` smooth '' of a terminal node .
we have distinguished above in section [ sect2 ] between : 1 . @xmath5 as representing an ultrametric set of relations , 2 . @xmath5 as representing an embedded set of sets . for case 1
we take @xmath260 as the @xmath3-dimensional observation vector corresponding to @xmath28 .
so , taking all @xmath11 vectors @xmath260 we have the initial data matrix @xmath261 of dimensions @xmath262 . then for our set of @xmath11 points in @xmath4 given in the form of matrix @xmath261 we have : @xmath263 where @xmath264 is the matrix collecting all wavelet projections or detail coefficients , @xmath21 .
the dimensions of @xmath172 are : @xmath265 ( see definition ( [ eqn2 ] ) ) .
the dimensions of @xmath264 are @xmath266 .
if @xmath267 is the final data smooth , in the limit for very large @xmath11 a constant - valued @xmath3-component vector , then let @xmath268 be the @xmath262 matrix with @xmath267 repeated on each of the @xmath11 rows .
consider the @xmath162th coordinate of the @xmath3-dimensional observation vector corresponding to @xmath28 .
for any @xmath269 we have : @xmath270 , i.e. the detail coefficient vectors are each of zero mean .
to recapitulate we have : @xmath261 is of dimensions @xmath262 .
@xmath172 is of dimensions @xmath271 .
@xmath264 is of dimensions @xmath272 .
@xmath268 is of dimensions @xmath273 .
we have distinguished between 1 . @xmath5 as representing an ultrametric set of relations , 2 . @xmath5 as representing an embedded set of sets .
we now turn attention to the latter . in this case
we take @xmath260 as an @xmath11-dimensional indicator vector corresponding to @xmath28 .
so , taking all @xmath11 vectors @xmath260 we have the initial data matrix @xmath261 which is none other than the @xmath274 dimensional identity matrix
. we will write @xmath275 for this identity matrix .
the wavelet transform in this case is : @xmath276 .
@xmath275 is of dimensions @xmath277 .
@xmath172 , exactly as in case 1 ( ultrametric case ) is of dimensions @xmath278 .
@xmath264 , of necessity different in values from case 1 , is of dimensions @xmath279 .
@xmath268 , of necessity different in values from case 1 , is of dimensions @xmath277 . in both cases considered ( viz . ,
ultrametric , and set of sets ) the forward and inverse transforms are performed in the same way .
the algorithms are identical the inputs alone differ .
the inverse transform allows exact reconstruction of the input data .
we begin with @xmath267 .
if this root node has subnodes @xmath12 and @xmath13 , we use @xmath280 and @xmath281 to form @xmath282 and @xmath283 . setting wavelet coefficients to zero and then reconstructing the data is referred to as hard thresholding ( in wavelet space ) and this is also termed wavelet smoothing or regression . see the companion paper , paper i , for discussion and examples .
we will look at the ultrametric case .
the matrix generalization of equation ( [ eqn3 ] ) is : @xmath284 matrix @xmath285 is formed from the vectors @xmath286 of equation ( [ eqn3 ] ) by replicating rows . now
the wavelet transform gives us : @xmath287 .
each ( replicated ) row of matrix @xmath268 is a particular measure of central tendency .
centering @xmath261 relative to this gives : @xmath288 we conclude from the formal similarity of expressions ( [ e612 ] ) and ( [ e613 ] ) : the initial p - adic encoding of our data vectors has been mapped into a _ wavelet encoding _ by the wavelet transform .
with reference to section [ dengroup ] , we note that relation [ e613 ] furnishes a unique matrix representation of a dendrogram .
generalization to regular @xmath0-way trees , for @xmath289 , may also be considered . for @xmath155 a natural wavelet function
is derived from the triangle scaling ( starck et al . , 1998 )
function , which is itself a convolution of a box function ( the scaling function defining the haar transform , used in this article ) with itself .
the haar scaling function used above was @xmath290 .
convolving this with itself gives then the scaling function @xmath291 .
convolving the box function again with the triangle function gives the b@xmath292 spline scaling function , @xmath293 , which is particularly natural for the analysis of a 5-way , @xmath294 , tree .
a remark on implementation follows : the 3-way tree is unfolded at each node into two 2-way trees .
more generally any regular @xmath0-way tree is unfolded at each node into @xmath184 two - way branchings .
the wavelet transform algorithm described previously is then directly applied .
we now look at other related work . in khrennikov and kozyrev ( 2004 ) and kozyrev ( 2001 )
the haar wavelet transform , defined on binary trees , was also introduced and discussed .
compared to the notation used here , the descriptions are related though a p - adic change of variable ( viz .
, @xmath295 is mapped onto @xmath296 . ) for the @xmath1 case , a convenient notational expression is given by the vladimirov operator ( see avetisov , bikulov , kozyrev and osipov , 2002 ) which is a modified differentiation operator .
the vladimirov operator is a p - adically expressed derivative for an ultrametric space with _ linearly _ related hierarchical levels , @xmath9 . in kozyrev ( 2001 , 2003 ) it is shown how the eigenvalues of the vladimirov operator are the haar wavelets . as a consequence ,
the hierarchical haar wavelet transform is a spectral analysis of the vladimorov operator .
our work differs from the works cited in the following way .
firstly , these other works deal with regular @xmath0-way trees .
degeneracies are allowed , which can cater for the irregular @xmath0-way trees that we have considered .
we have preferred to directly address the dendrogram data structure , given that it models observed data well .
secondly , these other works cater for infinite trees .
we have restricted ourselves to a more curtailed problem , with the aim of having a straightforward implementation , and with the aim of targeting the analysis of practical , constructive data analysis problems .
we have also been more focused in this work compared to the general setting described in comprehensive depth by benedetto and benedetto ( 2004 ) .
an important reason for considering dendrograms rather than infinite regular trees is that the former setting gives rise to ( low order ) polynomially bound algorithms for all operations ; whereas the latter , in the general case , are not polynomially bound .
a final path for future work will be noted .
the haar wavelet transform on a dendrogram ( @xmath5 ) gives us information on the rate of change of the clusters ( @xmath12 ) , with respect to the level index of each cluster ( @xmath9 ) . in a sense
this haar wavelet transform is the derivative of @xmath5 with respect to @xmath9 .
this perspective may be of benefit when dealing with the dynamics of ultrametric spaces ( avetisov et al .
, 2002 , and references therein ; kuhlmann , 2002 )
. 99 m.v .
altaisky , `` p - adic wavelet transform and quantum physics '' , proc .
steklov institute of mathematics , vol .
245 , 3439 , 2004 .
altaisky , m.v .
_ wavelets : theory , applications , implementation _ , universities press .
antoine , y.b .
kouagou , d. lambert and b. torrsani , `` an algebraic approach to discrete dilations .
application to discrete wavelet transforms '' , journal of fourier analysis and applications , 6 , 113141 , 2000 .
avetisov , v.a . ,
bikulov , a.h .
, kozyrev , s.v . and osipov , v.a .
`` p - adic models of ultrametric diffusion constrained by hierarchical energy landscapes '' , _ journal of physics a : mathematical and general _ , 35 , 177189 . bartal , y. , linial , n. , mendel , m. and naor , a. ( 2004 ) .
`` low dimensional embeddings of ultrametrics '' , _
european journal of combinatorics _
, 25 , 8792 .
benedetto , r.l .
`` examples of wavelets for local fields '' , in c. heil , p. jorgensen , d. larson , eds . , _ wavelets , frames , and operator theory _ , _ contemporary mathematics vol . 345 _ , 2747 .
benedetto , j.j . and
benedetto , r.l .
( 2004 ) . `` a wavelet theory for local fields and related groups '' , _ the journal of geometric analysis _
, 14 , 423456 .
benzcri , j.p .
( 1979 ) . _
la taxinomie _ , 2nd ed .
, paris : dunod .
benzcri , j.p . , translated by gopalan , t.k . ( 1992 ) .
_ correspondence analysis handbook _
, basel : marcel dekker .
k. bouchki , `` dcomposition des mesures et fonctions sur un ensemble fini probabilis muni dune classification arborescente .
i. arbres et hirarchies de parties '' , les cahiers de lanalyse des donns , xxi , no .
2 , 243254 , 1995 .
l. brekke and p.g.o .
freund , `` p - adic numbers in physics '' , _ physics reports _ , vol .
233 , 1993 , pp .
h. cendra and j.e .
marsden ( 2005 ) , geometric mechanics and the dynamics of asteroid pairs , dynamical systems , an international journal , 20 , 321 .
chakrabarti , k. , garofalakis , m. , rastogi , r. and shim , k. ( 2001 ) .
`` approximate query processing using wavelets '' , _ vldb journal , international journal on very large databases _
, 10 , 199223 .
chakraborty , p. ( 2005 ) .
`` looking through newly to the amazing irrationals '' , arxiv : math.ho/0502049v1 , 2 feb .
debnath , l. and mikusiski , p. ( 1999 ) .
_ introduction to hilbert spaces with applications _ , 2nd edn .
, academic press .
k. flornes , a. grossmann , m. holschneider and b. torrsani , `` wavelets on discrete fields '' , applied and computational harmonic analysis , 1 , 137147 , 1994 .
r. foote , g. mirchandani , d. rockmore , d. healy and t. olson `` a wreath product group approach to signal and image processing : part i multiresolution analysis '' , ieee trans . in signal processing ,
48(1 ) , 2000a , pp .
102132 r. foote , g. mirchandani , d. rockmore , d. healy and t. olson `` a wreath product group approach to signal and image processing : part ii
convolution , correlations and applications '' , ieee trans . in signal processing ,
48(3 ) , 2000b , pp . 749767 .
r. foote , `` an algebraic approach to multiresolution analysis '' , transactions of the american mathematical society , 357 , 50315050 , 2005 .
frazier , m.w .
( 1999 ) . _
an introduction to wavelets through linear algebra _ new york : springer .
gaji , l. ( 2001 ) .
`` on ultrametric space '' , _ novi sad journal of mathematics _ , 31 , 6971 .
gouva , f.q .
( 2003 ) . _ p - adic numbers _ , berlin : springer .
johnson , s.c .
`` hierarchical clustering schemes '' , _ psychometrika _ , 32 , 241254 .
kargupta , h. and park , b .- h .
( 2004 ) . `` a fourier spectrum - based approach to represent decision trees for mining data streams in mobile environments '' , _ ieee transactions on knowledge and data engineering _ , 16 , 216229 .
khrennikov , a.yu . and kozyrev , s.v .
`` pseudo - differential operators on ultrametric spaces and ultrametric wavelets '' , http://arxiv.org/abs/math-ph/0412062 khrennikov , a.yu . and kozyref , s.v .
( 2006 ) , `` ultrametric random field '' , http://arxiv.org/abs/math.pr/0603584 a.w .
knapp , `` group representations and harmonic analysis from euler to langlands '' , notices of the american mathematical society , 43 , 537549 , 1996 .
kozyrev , s.v .
`` wavelet analysis as a p - adic spectral analysis '' , _ math .
_ , 66 , 367376 .
http://arxiv.org/abs/math-ph/0012019 kozyrev , s.v .
`` p - adic pseudo - differential operators and p - adic wavelets '' , _ theoretical and mathematical physics _
, 138 , 322332 .
http://arxiv.org/abs/math-ph/0303045 kuhlmann , f .- v .
( 2002 ) , `` maps on ultrametric spaces , hensel s lemma , and differential equations over valued fields '' , preprint , http://math.usask.ca/fvk/recpapa.htm lemin , a.j .
`` isometric embedding of ultrametric ( non - archimedean ) spaces in hilber space and lebesgue space '' , in _ p - adic functional analysis _
, ioannina , 2000 , vol .
222 of _ lecture notes in pure and appl .
_ , pp . 203218 , dekker
. m. krasner , `` nombres semi - rels et espaces ultramtriques '' , _ comptes - rendus de lacadmie des sciences , tome ii _
, vol . 219 , 1944 , pp . 433 .
lang ( 1998 ) , `` wavelet analysis on the cantor dyadic group '' , houston journal of mathematics , 24 , 533544 .
addendum , 24 , 757758 .
lerman , i.c .
_ classification et analyse ordinale des donnes _ paris : dunod .
r. mojena , `` hierarchical grouping methods and stopping rules : an evaluation '' , the computer journal , 20 , 359363 , 1977 .
f. murtagh ( 1984 ) , `` counting dendrograms : a survey '' , discrete applied mathematics , 7 , 191199 , 1984 .
murtagh , f. ( 1984 ) .
_ multidimensional clustering algorithms _ wrzburg : physica - verlag .
murtagh , f. ( 2006 ) .
`` the haar wavelet transform of a dendrogram '' , referred to as paper i here , _
journal of classification _ , forthcoming , 2007 .
http://arxiv.org/abs/cs.ir/0608107 reiter , h. and stegeman , j.d .
_ classical harmonic analysis and locally compact groups _ , 2nd edition , oxford : oxford university press .
( definition 4.1.16 , p. 131 . ) van rooij , a.c.m .
( 1978 ) . _
non - archimedean functional analysis _ , dekker .
schikhof , w.h .
_ ultrametric calculus
_ , cambridge : cambridge university press .
( chapters 18 , 19 , 20 , 21 . ) sloan , n.j.a .
( 2005 ) , sequence a000111 , the on - line encyclopedia of integer sequences , www.research.att.com/@xmath297njas/sequences smart project ( 2005 ) , algebraic theory of signal processing , + www.ece.cmu.edu/@xmath297smart starck , j.l . ,
murtagh , f. and bijaoui , a. ( 1998 ) . _ image and data analysis : the multiscale approach _ ,
cambridge : cambridge university press .
strang , g. and nguyen , t. ( 1996 ) .
_ wavelets and filter banks _ , wellesley - cambridge press .
b. torrsani , `` some remarks on wavelet representations and geometric aspects '' , in wavelets , theory , algorithms and applications , c.k .
chui , l. montefusco and l. puccio , eds . ,
world scientific , pp .
91115 , 1994 .
b. torrsani , `` time - frequency analysis , from geometry to signal processing '' , in contemporary problems in mathematical physics , proc . of copromath conference ( nov .
1999 ) , j. govaerts , n. hounkonnou and w.a .
lester , eds .
, world scientific , pp .
7496 , 2000 .
vitter , j.s . and wang , m. ( 1999 ) .
`` approximate computation of multidimensional aggregates of sparse data using wavelets '' , in _ proceedings of the acm sigmod international conference on management of data _
, 193204 .
ward , t. ( 1994 ) .
`` entropy of compact group automorphisms '' , http://www.mth.uea.ac.uk/@xmath297h720/lecture_notes .
classically , the haar wavelet function basis for analysis of @xmath60 is determined by inducing the basis from an @xmath3-dimensional pixel ( time step , voxel , etc . )
grid , @xmath62 .
basis functions of a space denoted by @xmath298 are defined from a scaling function @xmath299 as follows ( starck , murtagh and bijaoui , 1998 ) : @xmath300 the functions @xmath299 are all box functions , defined on the interval @xmath301 and are piecewise constant on @xmath302 subintervals .
we can approximate any function in spaces @xmath298 associated with basis functions @xmath303 , in a very fine manner for @xmath304 ( in this case of @xmath304 , all values ) , more crudely for @xmath305 and so on .
we consider the nesting of spaces , @xmath306 .
equation ( 1 ) directly leads to a dyadic analysis .
next we consider the orthogonal complement of @xmath305 in @xmath307 , and call it @xmath308 .
the basis functions for @xmath309 are derived from the haar wavelet .
we find @xmath310 this leads to the basis for @xmath307 as being equal to : the basis for @xmath305 together with the basis for @xmath308 . in practice
we use this finding like this : we write a given function in terms of basis functions in @xmath307 ; then we rewrite in terms of basis functions in @xmath305 and @xmath308 ; and then we rewrite the former to yield , overall , an expression in terms of basis functions in @xmath311 , @xmath312 and @xmath308 .
the wavelet parts provide the detail part , and the space @xmath311 provides the smooth part . for the definitions of scaling function and wavelet function in the case of the haar wavelet transform , proceeding from the given signal , the spaces @xmath298 are formed by averaging of pairs of adjacent values , and the spaces @xmath309 are formed by differencing of pairs of adjacent values .
proceeding in this direction , from the given signal , we see that application of the scaling or wavelet functions involves downsampling of the data .
the low - pass filter is a moving average .
the high - pass filter is a moving difference .
other low- and high - pass filters are alternatively used to yield other wavelet transforms .
a hierarchy , @xmath5 , is defined as a binary , rooted , unlabeled , node - ranked tree , also termed a dendrogram ( benzcri , 1979 ; johnson , 1967 ; lerman , 1981 ; murtagh , 1985 ) .
a hierarchy defines a set of embedded subsets of a given set , @xmath25 .
however these subsets are totally ordered by an index function @xmath9 , which is a stronger condition than the partial order required by the subset relation .
a bijection exists between a hierarchy and an ultrametric space .
let us show these equivalences between embedded subsets , hierarchy , and binary tree , through the constructive approach of inducing @xmath5 on a set @xmath25 .
hierarchical agglomeration on @xmath11 observation vectors , @xmath26 , involves a series of @xmath313 pairwise agglomerations of observations or clusters , with the following properties .
a hierarchy @xmath314 such that ( i ) @xmath315 , ( ii ) @xmath316 , and ( iii ) for each @xmath317 .
here we have denoted the power set of set @xmath25 by @xmath318 .
an indexed hierarchy is the pair @xmath319 where the positive function defined on @xmath5 , i.e. , @xmath320 , satisfies : @xmath321 if @xmath322 is a singleton ; and ( ii ) @xmath323 .
here we have denoted the positive reals , including 0 , by @xmath245 .
function @xmath9 is the agglomeration level .
take @xmath324 , let @xmath325 and @xmath326 , and let @xmath14 be the lowest level cluster for which this is true .
then if we define @xmath327 , @xmath264 is an ultrametric . in practice ,
we start with a euclidean or other dissimilarity , use some criterion such as minimizing the change in variance resulting from the agglomerations , and then define @xmath328 as the dissimilarity associated with the agglomeration carried out .
p - adic numbers were introduced by kurt hensel in 1898 .
the ultrametric topology was introduced by marc krasner ( 1944 ) , the ultrametric inequality having been formulated by hausdorff in 1934 .
essential motivation for the study of this area is provided by schikhof ( 1984 ) as follows .
real and complex fields gave rise to the idea of studying any field @xmath329 with a complete valuation @xmath330 comparable to the absolute value function .
such fields satisfy the `` strong triangle inequality '' @xmath331 .
various terms are used interchangeably for analysis in and over such fields such as p - adic , ultrametric , non - archimedean , and isosceles .
the natural geometric ordering of metric valuations is on the real line , whereas in the ultrametric case the natural ordering is a hierarchical tree .
p - adic numbers , which provide an analytic version of ultrametric topologies , have a crucially important property resulting from ostrowski s theorem : each non - trivial valuation on the field of the rational numbers is equivalent either to the absolute value function or to some p - adic valuation ( schikhof , 1984 , p. 22 ) .
essentially this theorem states that the rationals can be expressed in terms of ( continuous ) reals , or ( discrete ) p - adic numbers , and no other alternative system .
the p - adic numbers are base p numbers , where p is a prime number .
it can be shown that the reals can be expressed as p - adic numbers where p is infinity .
the question then arises as to whether any one of p = 2 , 3 , 5 , 7 , 11 , @xmath332 can be preferred . for want of justification to limit attention to one or a few values of p , taking them all gives rise to the adelic number system ( brekke and freund , 1993 ) .
see elsewhere for the basic ultrametric inequality , and the triangle propert
isosceles with small base or equilateral .
the following is based on lerman ( 1981 ) , chapter 0 , part iv .
theorem 1 : every point of a circle in an ultrametric space is a center of the circle . proof 1 : it suffices to consider the triangle @xmath333 , where @xmath54 is the center of the given circle , @xmath334 is an element of this circle , and @xmath144 is an element of the circle with the same radius but with center @xmath334 .
this triangle is isosceles . from the triangle property
the result follows .
corollary 1 : two circles of the same radius , that are not disjoint , are overlapping . definition 1 : a divisor of the ultrametric space , @xmath335 , is an equivalence relation @xmath264 satisfying @xmath336 . corollary 2 : circles of the same radius form a partition of the ultrametric set .
the corresponding equivalence is a divisor of the space .
definition 2 : a valuation of a divisor @xmath264 of the space @xmath335 is the number @xmath337 .
corollary 3 : if @xmath264 and @xmath338 are two divisors in @xmath335 , a finite metric space , verifying @xmath339 , then @xmath340 and reciprocally .
theorem 2 : if @xmath172 adn @xmath341 are disjoint circles in @xmath335 , the distance @xmath342 of an @xmath343 and of an @xmath344 depends on @xmath172 and @xmath341 only , and not on @xmath144 and @xmath345 . proof 2 : consider the triangles @xmath346 , where @xmath347 and apply the ultrametric triangle relationship .
corollary 4 : the quotient @xmath348 of an ultrametric space by a divisor is an ultrametric space .
the distance between two of its points ist strictly greater than @xmath349 in the finite case
. definition 3 : an ultrametric proximity is a positive ( possibly infinite ) function @xmath350 , verifying ( i ) @xmath351 , ( ii ) @xmath352 iff @xmath353 ; and ( iii ) @xmath354 .
corollary 5 : if @xmath21 is an ultrametric distance , then @xmath355 is an ultrametric proximity . if @xmath0 is an ultrametric proximity , then @xmath356 is an ultrametric distance .
theorem 3 : for an @xmath274 matrix of positive reals , symmetric with respect to the principal diagonal , to be a matrix of distances associated with an ultrametric distance on @xmath335 , a sufficient and necessary condition is that a permutation of rows and columns satisfies the following form of the matrix : 1 . above the diagonal term , equal to 0 , the elements of the same row are non - decreasing .
2 . for every index @xmath357 , if @xmath358 then @xmath359 and @xmath360 under these circumstances , @xmath361 is the length of the section beginning , beyond the principal diagonal , the interval of columns of equal terms in row @xmath357 .
proof 4 : follows from ultrametric triangle inequality .
see lerman ( 1981 ) , p. 50 .
theorem 5 : in an ultrametric topology , every ball is both open and closed ( termed clopen ) . ( the empty set and the universal set are both clopen .
the complement of a clopen set is clopen .
finite unions and intersections of clopen sets are clopen . ) from chakraborty ( 2004 ) : a _ basic neighborhood _ of @xmath144 , of radius @xmath204 , is the set @xmath362 . an _
open set _ ,
@xmath363 , is a union of basic neighborhoods , i.e. @xmath364 s.t . @xmath365 .
many sets are open and closed at the same time .
this property is relative to subspaces .
let @xmath366 be a metric space and @xmath367 .
if @xmath368 and @xmath369 , let @xmath370 denote the basic neighborhood of @xmath345 in @xmath371 , and @xmath372 the basic neighborhood of @xmath345 in @xmath261 . then @xmath373 .
it follow that a set @xmath374 is open in @xmath371 iff @xmath375 an open set @xmath376 in @xmath261 s.t .
an analogous statement holds for closed sets .
if @xmath378 then @xmath379 can be open ( or closed ) in @xmath371 without being open ( or closed ) in @xmath261 .
informally , a set of points is of necessity 0-dimensional . van rooij ( 1978 ) : a topology is 0-dimensional if it has a base consisting of clopen sets .
i.e. , if for every @xmath385 and for every closed @xmath386 that does not contain @xmath54 , there exists a clopen set @xmath379 such that @xmath387 . | we consider the wavelet transform of a finite , rooted , node - ranked , @xmath0-way tree , focusing on the case of binary ( @xmath1 ) trees .
we study a haar wavelet transform on this tree .
wavelet transforms allow for multiresolution analysis through translation and dilation of a wavelet function .
we explore how this works in our tree context .
* keywords : * haar wavelet transform ; binary tree ; ultrametric topology ; p - adic numbers ; hierarchical clustering ; data mining ; local field ; abelian group . |
nonclassicality of optical fields has been a topic of great interest in quantum optics and quantum information processing @xcite , which is usually associated with quantum interference and entanglement .
the phase space wigner function ( wf ) @xcite of quantum states of light is a powerful tool for investigating such nonclassical effects .
the wf was first introduced by wigner in 1932 to calculate quantum corrections to a classical distribution function of a quantum - mechanical system .
the partial negativity of the wf is indeed a good indication of the highly nonclassical character of the state @xcite and monitors a decoherence process of a quantum state , e. g. the excited coherent state in both photon - loss and thermal channels @xcite , the single - photon subtracted squeezed vacuum state in both amplitude decay and phase damping channels @xcite , and so on r5,r6,r7,r8,r8a .
nonlinear interaction of light in a medium provides a very useful framework to study various nonclassical properties of quantum states of radiation .
the kerr medium is one of the simplest nonlinearity , which shall allow us to investigate the full time - dependent wf dynamics with or without a quantum noise .
recently , a fokker - planck equation for the wf evolution in a noisy kerr medium ( @xmath0 nonlinearity ) is presented @xcite .
then the authors numerically solved this equation assuming coherent state as an initial condition and discussed its dissipation effects . however , for any initial condition , as far as we are concerned , there is no report about the wf evolution . on the other hand , by using the thermo entangled state representation ( tesr ) we solved various master equations to obtain density operators with an infinite operator - sum representation @xcite and then revealed that the wf of density operator can be expressed as an overlap between two pure states ( see eq .
( [ 2.8 ] ) below ) @xcite .
this brings much convenience to calculate time evolution of wfs when quantum decoherence happens .
thus the tesr is beneficial to quantum decoherence theory . in this paper
, we shall appeal the tesr @xmath1 to treat the wf evolution at any initial condition in self - kerr medium with photon loss and present a new formula to calculate time evolution of the wf for quantum decoherence .
in addition , based on the derived wf evolution formula , we shall deduce the photon number distribution for any initial state in presence of kerr interaction , where the photon number distribution is independent of the coupling factor @xmath2 that is relative to the kerr medium . as examples ,
the wf formula is applied to the cases of initial coherent state , and number state , respectively .
conclusions are involved in the last section .
we begin with briefly reviewing the thermo entangled state representation ( tesr ) . on the basis of umezawa - takahash thermo field dynamics ( tfd ) 5,6,7
we constructed the tesr in the doubled fock space @xcite , @xmath3 \left\vert 0,\tilde { 0}% \right\rangle , \label{2.1}\end{aligned}\ ] ] where @xmath4 is the displacement operator , @xmath5 is a fictitious mode accompanying the real photon creation operator @xmath6 @xmath7 and @xmath8 is annihilated by @xmath9 with the relations @xmath10 = 1 $ ] and @xmath11 = 0 $ ] .
the structure of @xmath12 is similar to that of the epr eigenstateshown in ref .
operating @xmath13 and @xmath14 on @xmath15 in eq.([2.1 ] ) we can obtain the eigen - equations of @xmath12,@xmath16 note that @xmath17 = 0,$ ] thus @xmath1 is the common eigenvector of @xmath18 and @xmath19 using the normally ordered form of vacuum projector @xmath20 and the technique of integration within an ordered product ( iwop ) of operators @xcite , we can easily prove that @xmath21 is complete and orthonormal,@xmath22 it is easily seen that @xmath23 has the properties@xmath24 ( where @xmath25 , and @xmath26 denotes the number in the fictitious hilbert space ) and@xmath27 note that density operators @xmath28,@xmath29 are defined in the real space which are commutative with operators ( @xmath5,@xmath30 in the tilde space . in a similar way , we can introduce the state vector @xmath31 conjugated to @xmath1 , defined as @xmath32 which also possesses orthonormal and complete properties@xmath33
in the markov approximation and interaction picture the master equation for a dissipative cavity with kerr medium has the form @xcite@xmath34 + \gamma \left ( 2a\rho a^{\dagger } -a^{\dagger } a\rho -\rho a^{\dagger } a\right ) , \label{3.5}\]]where @xmath35 is decaying parameter of the dissipative cavity , @xmath36 is coupling factor depending on the kerr medium .
milburn and holmes @xcite solved this equation by changing it to a partial differential equation for the q - function and for an initial coherent state . here
we will solve the master equation by virtue of the entangled state representation and present the infinite sum representation of density operator .
operating the both sides of eq.([3.5 ] ) on the state @xmath37 letting @xmath38 ( here one should understand the single - mode density operator @xmath39 in the left of eq.([3.5 ] ) as the direct product @xmath40 when @xmath39 acts onto the two - mode state @xmath41 , where @xmath42 is the identity operator in the auxiliary mode . ) , and using eq.([2.5 ] ) we can convert the master equation in eq .
( [ 3.5 ] ) into the following form , @xmath43 \right .
\notag \\ & & + \left .
\gamma \left ( 2a\tilde{a}-a^{\dagger } a-\tilde{a}^{\dagger } \tilde{%
a}\right ) \right\ } \left\vert \rho \right\rangle , \label{3.6}\end{aligned}\ ] ] i.e. , an evolution equation of state vector @xmath44 .
its solution is then of the form@xmath45 + \gamma t\left ( 2a\tilde{a}-a^{\dagger } a-\tilde{a}^{\dagger } \tilde{a}\right ) } \left\vert \rho _ { 0}\right\rangle , \label{3.7}\]]where @xmath46 @xmath47 is an initial density operator .
the advantage of using thermo field notation over more traditional algebraic manipulation with superoperators is that in many situations ( and , particularly , ones of our interest ) it enables to simplify , make more illustrative and less cumbersome finding the solution ( [ 3.7 ] ) and estimation of time - dependent matrix elements . in particular
, it allows to represent in a simple form a factorization of the superoperator exp\{@xmath48 } into multipliers with easily estimated actions on the number states @xcite . by introducing the following operators , @xmath49which satisfy @xmath50 = \left [ k_{0},k_{-}\right ] = 0,$ ] we can rewrite eq.([3.7 ] ) as@xmath51 + \gamma t\left ( 2k_{-}-2k_{z}+1\right ) \right\ } }
\left\vert \rho _ { 0}\right\rangle \notag \\ & = \exp \left [ i\chi tk_{0}+\gamma t\right ] \notag \\ & \times \exp \left\ { -2t\left ( \gamma + i\chi k_{0}\right ) \left [ k_{z}+% \frac{-\gamma } { \gamma + i\chi k_{0}}k_{-}\right ] \right\ } \left\vert \rho _ { 0}\right\rangle . \label{3.9}\end{aligned}\]]with the aid of the operator identity @xcite@xmath52 \notag \\ & = & \exp \left [ \sigma b\left ( e^{\lambda \tau } -1\right ) /\tau \right ] e^{\lambda a } , \label{3.10}\end{aligned}\]]which is valid for @xmath53 = \tau b,$ ] and noticing @xmath54 = -k_{-},$ ] we can reform eq.([3.9 ] ) as@xmath55 \exp % \left [ \gamma _ { z}k_{z}\right ] \exp \left [ \gamma _ { -}k_{-}\right ] \left\vert \rho _ { 0}\right\rangle , \label{2.12}\]]where@xmath56from eq.([2.12 ] ) we can obtain the infinite operator - sum form of @xmath57 , ( see appendix a ) @xmath58where the two operators @xmath59 and @xmath60 are respectively defined as@xmath61although @xmath59 is not hermite conjugate to @xmath62 , the normalization still holds , ( @xmath63 see appendix b ijtp ) i.e. , they are trace - preserving in a general sense , so @xmath59 and @xmath60 may be named the generalized kraus operators .
in this section , we consider wigner function s time evolution in the self - kerr medium channel . for this purpose
, we shall derive a new expression of wigner function in the tesr . according to the definition of wigner function of density operator @xmath64 @xmath65 , \label{2.6}\ ] ] where @xmath66 is the single - mode wigner operator @xcite
, whose explicit normally ordered form is @xcite @xmath67 by using @xmath68 ( @xmath69 ) and noticing ( [ 2.9 ] ) as well as @xmath70 * * * * we can reform eq.([2.6 ] ) as@xmath71 where eq .
( [ 2.8 ] ) is the wigner function formula in thermo entangled state representation , with which the wigner function of density operator is simplified as an overlap between two pure states in enlarged fock space , rather than using ensemble average in the system - mode space .
this will brings much convenience to calculate the time evolution of wigner functions when quantum decoherence happens . projecting ( [ 2.12 ] ) on the entangled state representation @xmath72 and inserting the completeness relation ( [ 2.10 ] ) , we find @xmath73where @xmath74 and @xmath75 are the wigner functions at the evolving time @xmath76 and initial time , respectively , and @xmath77 \notag \\ & & \times \exp \left [ \gamma _ { z}k_{z}\right ] \exp \left [ \gamma _ { -}k_{-}% \right ] \left\vert \xi _ { = 2\beta } ^{\prime } \right\rangle .
\label{3.2}\end{aligned}\]]it is convenient to the matrix element in ( [ 3.2 ] ) according to the two - mode fock space .
thus the @xmath78 is expanded as @xmath79by using the two - mode fock state @xmath80 , we get @xmath81where @xmath82 is the two - variable hermite polynomials @xcite . inserting the complete relation @xmath83 after a long but straight calculation , then the wigner function s evolution is given by ( see appendix c)@xmath84where @xmath85@xmath86and@xmath87it is obvious that , when @xmath88 the case of photon loss , @xmath89 and eq.([3.11 ] ) just does reduce to ( see appendix e ) @xmath90which is just the evolving formula of wigner function for amplitude - damping channel . while for @xmath91 without photon - loss , eq.([3.11 ] ) reduces to @xmath92 } { m!n!e^{2\left\vert \alpha \right\vert ^{2}}}h_{m , n}\left ( 2\alpha ^{\ast } , 2\alpha \right ) \notag \\ & \times 4\int \frac{\mathtt{d}^{2}\beta } { \pi } e^{-2\left\vert \beta \right\vert ^{2}}h_{m , n}\left ( 2\beta , 2\beta ^{\ast } \right ) w\left ( \beta , \beta ^{\ast } , 0\right ) .
\label{3.15}\end{aligned}\ ] ]
now we consider photon number ( pn ) distribution in presence of kerr medium . according to the tfd
, we can reform the pn @xmath93 $ ] as @xmath94which is converted to the matrix element @xmath95 in the context of thermo dynamics . then using the completeness of @xmath96 and eq.(5.1 ) as well as eq.([2.8 ] ) , we have@xmath97where @xmath98 is the wigner function of number state @xmath99 as shown in @xcite .
thus one can calculate the pn by combining eqs.([3.11 ] ) and ( [ 5.2 ] ) .
next we evaluate the pn for the above decoherence model in eq.([3.5 ] ) . substituting eq.([3.11 ] ) into eq.([5.2 ] ) , we have @xmath100where @xmath101then substituting eqs.([5.4 ] ) and ( [ 3.13 ] ) into ( [ 5.3 ] ) yields @xmath102which corresponds to the photon number of density operator in the amplitude - damping quantum channel @xcite . from eq.([5.5 ] ) , it is easily to see that , for any initial state , the photon number distribution @xmath103 is independent of the coupling factor @xmath36 that is relative to the kerr medium , as respected in @xcite .
the phase space wigner distribution function description of quantum states of light is a powerful tool to investigate nonclassical effects , such as quantum interference and entanglement . in this section , as the applications of wf evolution formula , we take two special initial states as examples .
\(1 ) when the initial state is the coherent state @xmath104 , whose wf is given @xmath105 thus substituting into eq.(3.13 ) yields ( see appendix g)@xmath106so@xmath107which is a new expression of the evolution of wf for any initial state . in particular , when @xmath91 without the dissipation , eq.([5.7 ] ) reduces to @xmath108which is identical to eq.(7 ) in ref.@xcite , where eq.(7 ) is used to make the numerical calculation since it is much more rapid than the other expression eq.(6 ) .
in addition , from eq.([5.7 ] ) one can see that the wf can be obtained very quickly when the dissipation can not be negligible .
further when @xmath109 , eq.([5.8 ] ) just returns to the wf of the initial coherent state .
fig.1 presents the plots of the wf for different parameters @xmath110 and @xmath111 from fig.1 , one can see that the wf turns into an ellipse and squeezing appears in an appropriate direction .
then the ellipse changes into a banana shape and a tailor of the interference fringes appears where the distribution takes the negative values .
[ fig8.1 ] and @xmath112 @xmath113 @xmath114
@xmath115 @xmath116 @xmath117 @xmath118,title="fig:",width=288 ] \(2 ) another example is number state , where the wfof number state @xmath119 is given by @xmath120substituting it into ( [ 3.13 ] ) and using the generating function of two - variable hermite polynomials [ see below eq.(f3 ) ] , we have@xmath121thus the evolution of wf for @xmath122 is@xmath123which is the wf of number state in the photon - loss channel and indicates that the number state is not affected by the kerr nonlinearity . in particular , when @xmath91 or @xmath124 eq.([5.11 ] ) just reduces to the wf of number state . from eq.([5.11 ] ) , on the other hand , it is found that the number state evolves into a density operator of binomial distribution ( a mixed state ) if @xmath125 is a binomial parameter .
in summary , by converting wigner function for quantum state into an overlap between two `` pure states '' in a two - mode enlarged fock space , we investigate the wf evolution of any initial condition in self - kerr medium with photon loss and present a new formula for calculating time evolution of the wf for quantum decoherence . based on the derived wf evolution formula , in addition
, we discuss the photon number distribution for any initial state in presence of kerr interaction .
it is found that the photon number distribution is independent of the coupling factor @xmath36 in correlation with the kerr medium , as expected by people . as applications , furthermore
, the two cases of initial coherent state and number state are considered .
it is shown that the coherent state can be squeezed due to the presence of kerr medium , while the number state is not affected by the kerr nonlinearity and evolves into a density operator of binomial distribution ( a mixed state ) with @xmath126 being a binomial parameter .
* acknowledgements * work supported by a grant from the key programs foundation of ministry of education of china ( no .
210115 ) and the research foundation of the education department of jiangxi province of china ( no .
gjj10097 ) . in order to obtain the infinite operator - sum form of @xmath127 from eq.([2.12 ] ) , using the completeness relation of fock state in the enlarged space @xmath128 and noticing @xmath129 , we have@xmath130where @xmath131furthermore , using the relations @xmath132we find @xmath133thus eq.(a1 ) becomes @xmath134where @xmath135 using eq.(a3 ) again , we see@xmath136after depriving @xmath137 from the both sides of eq.(a6 ) , the solution of master equation ( [ 3.5 ] ) appears as an infinite operator - sum form note that the factor @xmath139 appears in the denominator of @xmath140 ( see eq.(a2 ) ) , ( this is originated from the nonlinear term of @xmath141 ) so that it is impossible to move all @xmath142dependent@xmath143terms to the right of @xmath144 .
fortunately , we can formally express eq.(a7 ) as eq.(2.26 ) .
using the operator identity @xmath145 \colon$ ] and the iwop technique , we can prove that @xmath146 \left ( a^{\dagger } a\right ) ^{l}\colon=1 , \tag{b1}\end{aligned}\ ] ] from which one can see that the normalization still holds , i.e. , they are trace - preserving in a general sense , so @xmath59 and @xmath147 may be named the generalized kraus operators . further using a new sum formula
( see appendix d ) @xmath149thus eq.(c1 ) can be recast into the following form@xmath150where @xmath151 is defined in ( 3.12 ) . substituting eq.(c3 ) into ( [ 3.1 ] ) yields ( [ 3.11 ] ) and ( 3.13 ) .
@xmath153 \notag \\ & = e^{xy}(-1)^{n}\int \frac{d^{2}z}{\pi } \sum_{l=0}^{\infty } \frac{\left ( -\alpha \left\vert z\right\vert ^{2}\right ) ^{l}}{l!}z^{n}z^{\ast m}\exp % \left [ -\left\vert z\right\vert ^{2}+xz - yz^{\ast } \right ] \notag \\ & = e^{xy}(-1)^{n}\int \frac{d^{2}z}{\pi } z^{n}z^{\ast m}\exp \left [ -\left ( \alpha + 1\right ) \left\vert z\right\vert ^{2}+xz - yz^{\ast } \right ] \notag \\ & = \frac{e^{\frac{\alpha \allowbreak xy}{\alpha + 1}}}{\left ( \alpha + 1\right ) ^{(m+n+2)/2}}(-1)^{n}e^{xy/\left ( \alpha + 1\right ) } \int \frac{% d^{2}z}{\pi } z^{n}z^{\ast m}\exp \left [ -\left\vert z\right\vert ^{2}+\frac{% xz - yz^{\ast } } { \sqrt{\alpha + 1}}\right ] \notag \\ & = \frac{e^{\frac{\alpha \allowbreak xy}{\alpha + 1}}}{\left ( \alpha + 1\right ) ^{(m+n+2)/2}}h_{m , n}\left ( \frac{x}{\sqrt{\alpha + 1}},\frac{y}{% \sqrt{\alpha + 1}}\right ) , \tag{d2}\end{aligned}\ ] ] when @xmath88 @xmath154 and @xmath155 } { m!n!\left ( t+1\right ) ^{(m+n+2)/2}}h_{m , n}\left ( 2\alpha ^{\ast } , 2\alpha \right ) e^{-2\left\vert \alpha \right\vert ^{2 } } , \tag{e1}\end{aligned}\]]then we have @xmath156using the following formula@xmath157 , \tag{e3}\end{aligned}\]]eq.(e2 ) can be rewritten as @xmath158thus eq.([3.11 ] ) becomes using the relation between hermite polynomial and lagurre polynomial,@xmath160 we can recast the left of eq.(*[5.4 ] * ) into the following form@xmath161 further using the generating function of @xmath162 , @xmath163 \right\vert _ { t = t^{\prime}=0 } , \tag{f3}\ ] ] and the integration formula,@xmath164 we have @xmath165 \notag \\ & \times\int\frac{\mathtt{d}^{2}\alpha}{\pi}\left .
\exp\left [ -\left\vert \alpha\right\vert ^{2}+\left ( t+\tau\right ) \alpha^{\ast}+\left ( t^{\prime } + \tau^{\prime}\right ) \alpha\right ] \right\vert _ { t = t^{\prime}=\tau = \tau^{\prime}=0 } \notag \\ & = \frac{\partial^{m^{\prime}+n^{\prime}}}{\partial t^{m^{\prime}}\partial t^{\prime n^{\prime}}}\frac{\partial^{m+n}}{\partial\tau^{m}\partial \tau^{\prime n}}\exp\left [ t\tau^{\prime}+\tau t^{\prime}\right ] _ { t = t^{\prime}=\tau=\tau^{\prime}=0 } \notag \\ & = m!n!\delta_{m^{\prime},n}\delta_{n^{\prime},m } , \tag{f5}\end{aligned}\ ] ] thus eq.(f2 ) becomes the right hand side of eq.(*[5.4 ] * ) .
@xmath168 \right\vert _ { t = t^{\prime}=0 } \notag \\ & = e^{-2\left\vert z\right\vert ^{2}}\frac{\partial^{m+n}}{\partial t^{m}\partial t^{\prime n}}e^{-tt^{\prime}}\int\frac{\mathtt{d}^{2}\beta}{% \pi^{2}}\left . \exp\left [ -\frac{y+1}{2}\left\vert \beta\right\vert ^{2}+\beta\left ( xt+z^{\ast}\right ) + \beta^{\ast}\left ( xt^{\prime } + z\right ) \right ] \right\vert _ { t = t^{\prime}=0 } \notag \\ & = \frac{1}{\pi}\frac{2}{y+1}e^{-\frac{2y}{y+1}\left\vert z\right\vert ^{2}}% \frac{\partial^{m+n}}{\partial t^{m}\partial t^{\prime n}}\exp\left [ -\left ( 1-\frac{2x^{2}}{y+1}\right ) tt^{\prime}+\frac{2xz}{y+1}t+\frac{2xz^{\ast}}{% y+1}t^{\prime}\right ] _ { t = t^{\prime}=0}. \tag{g3}\end{aligned}\ ] ] | by introducing the thermo entangled state representation , we convert the calculation of wigner function ( wf ) of density operator to an overlap between `` two pure '' states in a two - mode enlarged fock space .
furthermore , we derive a new wf evolution formula of any initial state in self - kerr medium with photon loss and find that the photon number distribution for any initial state is independent of the coupling factor with kerr medium , where the number state is not affected by the kerr nonlinearity and evolves into a density operator of binomial distribution .
* keywords : * wigner function , kerr medium , entangled state representation |
the broadcast nature of wireless transmissions causes co - channel interference and channel contention , which can be viewed as interactions among transceivers .
interactions among multiple decision makers can be formulated and analyzed using a branch of applied mathematics called game theory @xcite .
game - theoretic approaches have been applied to a wide range of wireless communication technologies , including transmission power control for code division multiple access ( cdma ) cellular systems @xcite and cognitive radios @xcite . for a summary of game - theoretic approaches to wireless networks ,
we refer the interested reader to @xcite .
application - specific surveys of cognitive radios and sensor networks can be found in @xcite . in this paper , we focus on potential games @xcite , which form a class of strategic form games with the following desirable properties : * the existence of a nash equilibrium in potential games is guaranteed in many practical situations @xcite ( theorems [ th : existence_finite ] and [ th : existence_infinite ] in this paper ) , but is not guaranteed for general strategic form games .
other classes of games possessing nash equilibria are summarized in @xcite and @xcite .
* unilateral improvement dynamics in potential games with finite strategy sets are guaranteed to converge to the nash equilibrium in a finite number of steps , i.e. , they do not cycle @xcite ( theorem [ eq : converge_in_finite_steps ] in this paper ) . as a result
, learning algorithms can be systematically designed .
a game that does not have these properties is discussed in example [ example2 ] in section [ sec : game ] .
we provide an overview of problems in wireless networks that can be formulated in terms of potential games .
we also clarify the relations among games , and provide simpler proofs of some known results .
problem - specific learning algorithms @xcite are beyond the scope of this paper . [ cols="<,<,<,<",options="header " , ] [ tab : system_models ] the remainder of this paper is organized as follows : in sections [ sec : game ] , [ sec : pot ] , and [ sec : learning ] , we introduce strategic form games , potential games , and learning algorithms , respectively .
we then discuss various potential games in sections [ ssec : nie ] to [ sec : immobile_sensor ] , as shown in table [ tab : system_models ] .
finally , we provide a few concluding remarks in section [ sec : conclusion ] .
the notation used here is shown in table [ 210446_20aug14 ] . unless the context indicates otherwise ,
sets of strategies are denoted by calligraphic uppercase letters , e.g. , @xmath0 , strategies are denoted by lowercase letters , e.g. , @xmath1 , and tuples of strategies are denoted by boldface lowercase letters , e.g. , @xmath2 .
note that @xmath3 is a scalar variable when @xmath0 is a set of scalars or indices , @xmath3 is a vector variable when @xmath0 is a set of vectors , and @xmath3 is a set variable when @xmath0 is a collection of sets .
we use @xmath4 to denote the set of real numbers , @xmath5 to denote the set of nonnegative real numbers , @xmath6 to denote the set of positive real numbers , and @xmath7 to denote the set of complex numbers .
the cardinality of set @xmath8 is denoted by @xmath9 .
the power set of @xmath8 is denoted by @xmath10 . finally , @xmath11 is the indicator function , which is one when @xmath12 is true and is zero otherwise .
we treat many system models , as shown in fig .
[ fig : system_models ] . in multiple - access channels , as shown in fig .
[ fig : system_models ] , multiple transmitters ( txs / users / mobile stations / terminals ) transmit signals to a single receiver ( rx / base station ( bs)/access point ( ap ) ) . in fig .
[ fig : system_models ] , @xmath13 represents the link gain from tx @xmath14 to the rx . in a network model consisting of tx - rx pairs , as shown in fig .
[ fig : system_models ] , each tx @xmath15 transmits signals to rx @xmath15 . in this case ,
@xmath16 . in a network model consisting of txs shown in fig .
[ fig : system_models ] , each tx ( bs / ap / transceiver / station / terminal / node ) interferes with others . in this model , @xmath17 . a `` canonical network model '' @xcite , shown in fig .
[ fig : system_models ] , consists of clusters that are spatially separated in order for @xmath18 to hold .
note that these network models have been discussed in terms of graph structure in @xcite .
we use @xmath19 to denote a directed link from tx @xmath15 to tx @xmath20 or cluster @xmath15 to cluster @xmath20 .
let interference graph @xmath21 be an undirected graph , where the set of vertices @xmath22 corresponds to txs or clusters , and @xmath15 interferes with @xmath20 if @xmath23 , as shown in fig .
[ fig : system_models ] , i.e. , @xmath24 where @xmath25 is the transmission power level for every tx and @xmath26 is a threshold of the received power .
note that in undirected graph @xmath21 , @xmath27 , for every @xmath28 .
we denote the neighborhood of @xmath14 in graph @xmath21 by @xmath29 .
we also define @xmath30 , then , @xmath31 .
c|x @xmath32 & strategic form game + @xmath33 & finite set of players , @xmath34 + @xmath35 & @xmath36 + @xmath37 & set of strategies for player @xmath38 + @xmath39 & strategy space , @xmath40 + @xmath41 & payoff function for player @xmath38 + @xmath42 & potential function + @xmath43 & best - response correspondence of player @xmath14 + @xmath3 & strategy of player @xmath15 , @xmath1 + @xmath44 & set of probability distributions over @xmath0 + @xmath45 & mixed strategy , @xmath46 + @xmath47 & mixed strategy profile , @xmath48 + @xmath49 & link gain between tx @xmath14 and a single isolated rx in fig .
[ fig : system_models ] + @xmath50 & link gain between tx @xmath51 and rx @xmath14 ; @xmath52 in fig .
[ fig : system_models ] , and @xmath53 in figs .
[ fig : system_models ] and [ fig : system_models ] + @xmath19 & directed link from @xmath15 to @xmath20 + @xmath54 & set of edges in undirected graph + @xmath55 & @xmath56 .
neighborhood in graph @xmath21 + @xmath57 & @xmath58
. + @xmath59 & common noise power for every player + @xmath60 & noise power at rx @xmath14 + @xmath61 & noise power at rx @xmath14 in channel @xmath62 + @xmath63 & interference power at rx @xmath14 at channel arrangement @xmath64 + @xmath65 & set of available channels for player @xmath14 + @xmath66 & channel of player @xmath14 + @xmath64 & @xmath67 + @xmath68 & set of available transmission power levels for player @xmath14 + @xmath69 & transmission power level of player @xmath14 as a strategy + @xmath70 & @xmath71 + @xmath25 & identical transmission power level for every player + @xmath72 & transmission power level for player @xmath15 as a constant + @xmath73 & required signal - to - interference - plus - noise power ratio ( sinr ) + [ 210446_20aug14 ]
we begin with the definition of a strategic form game and present an example of a game - theoretic formulation of a simple channel selection problem .
moreover , we discuss other useful concepts , such as the best response and nash equilibrium .
the analysis of nash equilibria in the channel selection example reveals the potential presence of cycles in best - response adjustments . a _ strategic ( or normal ) form game _ is a triplet @xmath74 , or simply @xmath75 , where @xmath34 is a finite set of _ players _ ( decision makers ) , @xmath76 is the set of _ strategies _ ( or actions ) for player @xmath77 , and @xmath78 is the _ payoff _ ( or utility ) function of player @xmath77 that must be maximized .
if @xmath79 , we denote the cartesian product @xmath80 by @xmath81 . if @xmath82 , we simply write @xmath83 to denote @xmath84 , and @xmath85 to denote @xmath86 . when @xmath87 , we let @xmath88 denote @xmath89 and @xmath90 denote @xmath91 . for @xmath92 , @xmath93 , @xmath94 , and @xmath95 . [ example1 ] consider a channel selection problem in the tx - rx pair model shown in fig . [
fig : system_models ] .
each tx - rx pair is assumed to select its channel in a decentralized manner in order to minimize the received interference power .
the channel selection problem can be formulated as a strategic form game @xmath96 .
the elements of the game are as follows : the set of players @xmath33 is the set of tx - rx pairs .
the strategy set for each pair @xmath15 , @xmath97 is the set of available channels .
the received interference power at rx @xmath38 is determined by a combination of channels @xmath98 , where @xmath99 let @xmath100 be the payoff function to be maximized , i.e. , @xmath101 note that @xmath102 was introduced in @xcite , and we further discuss it in example [ example2 ] .
the _ _ best - response correspondence _ _ ( or simply , best response ) @xmath103 of player @xmath14 to strategy profile @xmath104 is the correspondence @xmath105 or equivalently , @xmath106 .
a fundamental solution concept for strategic form games is the nash equilibrium : a strategy profile @xmath107 is a pure - strategy _ nash equilibrium _ ( or simply a nash equilibrium ) of game @xmath108 if @xmath109 for every @xmath38 and @xmath92 ; equivalently , @xmath110 for every @xmath38 .
that is , @xmath111 is a solution to the optimization problem @xmath112 . at the nash equilibrium
, no player can improve his / her payoff by adopting a different strategy _ unilaterally _ ; thus , no player has an incentive to unilaterally deviate from the equilibrium .
the nash equilibrium is a proper solution concept ; however , the existence of a pure - strategy nash equilibrium is not necessarily guaranteed , as shown in the next example . .
a cycle results from the best - response adjustment . ]
[ example2 ] consider @xmath102 and the arrangement shown in fig .
[ fig : loop ] , i.e. , @xmath113 , @xmath114 for every @xmath14 , and @xmath115 , @xmath116 , and @xmath117 .
the game does not have a nash equilibrium , i.e. , for every channel allocation , at least one pair has an incentive to change his / her channel .
the details are as follows : when all players choose the same channel , e.g. , @xmath118 , every player has an incentive to change his / her channel because @xmath119 for all @xmath14 ; thus , it is not in nash equilibrium . on the contrary , when two players choose the same channel , and the third player chooses a different channel , e.g. , @xmath120 , as shown in fig .
[ fig : loop](a ) , @xmath121 , i.e. , pair 2 has an incentive to change its channel @xmath122 from 1 to 2 , and ( [ eq : ne ] ) does not hold . because of the symmetry property of the arrangement in fig .
[ fig : loop ] , every strategy profile does not satisfy ( [ eq : ne ] ) .
furthermore , the best - response channel adjustments , which will be formally discussed in section [ sec : learning ] , cycle as @xmath123 , @xmath124 , @xmath125 , @xmath126 , @xmath127 , @xmath128 , and @xmath123 , as shown in figs .
[ fig : loop](a - f ) . the channel allocation game @xmath129 is discussed further in section [ ssec : nie ] .
we state key definitions and properties of potential games in section [ ssec : properties ] , show how to identify and design exact potential games in sections [ ssec : identification_exact ] and [ ssec : design ] , and show how to identify ordinal potential games in section [ ssec : identification_ordinal ] .
monderer and shapley @xcite introduced the following classes of potential games : a strategic form game @xmath130 is an _ exact potential game _ ( epg ) if there exists an _ exact potential function _
@xmath131 such that @xmath132 for every @xmath77 , @xmath133 , and @xmath134 .
a strategic form game @xmath135 is a _ weighted potential game _ ( wpg ) if there exist a _ weighted potential function _ @xmath131 and a set of positive numbers @xmath136 such that @xmath137 for every @xmath138 , @xmath133 , and @xmath134 .
a strategic form game @xmath135 is an _ ordinal potential game _ ( opg ) if there exists an _
ordinal potential function _
@xmath131 such that @xmath139 for every @xmath138 , @xmath133 , and @xmath134 , where @xmath140 denotes the sign function .
although the potential function @xmath42 is independent of the indices of the players , @xmath42 reflects any unilateral change in any payoff function @xmath141 for every player @xmath14 .
since an epg is a wpg and a wpg is an opg @xcite , the following properties of opgs are satisfied by epgs and wpgs .
[ th : existence_finite ] every opg with finite strategy sets possesses at least one nash equilibrium ( * ? ? ?
* corollary 2.2 ) .
[ th : existence_infinite ] in the case of infinite strategy sets , every opg with compact strategy sets and continuous payoff functions possesses at least one nash equilibrium ( * ? ? ?
* lemma 4.3 ) .
[ th : unique ] every opg with a compact and convex strategy space , and a strictly concave and continuously differentiable potential function possesses a unique nash equilibrium ( * ? ? ?
* theorem 2),@xcite .
the most important property of potential games is _ acyclicity _ , which is also referred to as the finite improvement property .
a _ path _ in @xmath142 is a sequence @xmath143 , { \boldsymbol a}[1 ] , \ldots)$ ] such that for every integer @xmath144 , there exists a unique player @xmath14 such that @xmath145 \neq a_i[k\!-\!1 ] \in \mathcal{a}_i$ ] while @xmath146 = { \boldsymbol a}_{-i}[k\!-\!1]$ ] .
@xmath143 , { \boldsymbol a}[1 ] , \ldots)$ ] is an _ improvement path _ if , for every @xmath147 , @xmath148 ) > u_i({\boldsymbol a}[k\!-\!1])$ ] , where @xmath14 is the unique deviator at step @xmath149 .
@xmath142 has the _ finite improvement property ( fip ) _ if every improvement path is finite .
[ eq : converge_in_finite_steps ] every opg with finite strategy sets has the fip ( * ? ? ?
* lemma 2.3 ) ; that is , unilateral improvement dynamics are guaranteed to converge to a nash equilibrium in a finite number of steps .
the definition of an epg utilizes a potential function ( [ eq : epg ] ) . sometimes , however , it is beneficial to know if a given game is an epg independently of its potential function .
the following properties of epgs and classes of games known to be epgs are useful for the identification and derivation of potential functions .
note that each epg has a unique exact potential function except for an additive constant ( * ? ? ?
* lemma 2.7 ) .
let @xmath150 be a strategic form game where strategy sets @xmath151 are intervals of real numbers and payoff functions @xmath41 are twice continuously differentiable .
then , the game is an epg if and only if @xmath152 for every @xmath153 ( * ? ? ?
* theorem 4.5 ) .
[ th : linear ] let @xmath154 and @xmath155 be epgs with potential functions @xmath156 and @xmath157 , respectively .
furthermore , let @xmath158 .
then , @xmath159 is an epg with potential function @xmath160 @xcite . if @xmath161 for all @xmath38 , where @xmath162 , the game @xmath163 is called a _ _ coordination game _ _ or an _ identical interest game _ , and @xmath164 is called a coordination function @xcite . if @xmath165 for all @xmath38 , where @xmath166 , the game @xmath167 is called a _ dummy game _ , and @xmath168 is called a dummy function @xcite . if @xmath169 for all @xmath38 , where @xmath170 , the game @xmath171 is called a _ self - motivated game _ , and @xmath172 is called a self - motivated function @xcite .
[ th : coord - dummy ] @xmath135 is an epg if and only if there exist functions @xmath173 and @xmath174 such that @xmath175 for every @xmath38 @xcite .
this game is said to be a _ coordination - dummy game_. the potential function of this game is @xmath176 . from theorem [ th : coord - dummy ] ,
any identical interest game is an epg .
almost all games found in studies applying identical interest games @xcite have the form of game @xmath177 , where @xmath178 for every @xmath38 and @xmath179 is a performance indicator of player @xmath20 , e.g. , @xmath180 is the individual throughput and @xmath181 is the aggregated throughput of all players @xcite . note that in most of these works , @xmath182 is used for comparison with other games .
closely related to @xmath182 , the form of game @xmath183 with payoff @xmath184 where @xmath185 , is found in many scenarios : data stream control in multiple - input and multiple - output ( mimo ) @xcite , channel assignment @xcite , joint power , channel and bs assignment @xcite , joint power and user scheduling @xcite , bs selection @xcite , and bs sleeping @xcite .
note that @xmath186 is not an identical interest game , but can be seen as @xmath182 on graphs , where the performance indicator of player @xmath15 is a function of strategies of its neighbors , i.e. , @xmath185 , and the sum of the performance indicators of player @xmath15 and neighbors @xmath187 is set for the payoff function of player @xmath15 .
it can be easily proved that @xmath186 is an epg with potential @xmath188 a strategic form game @xmath189 is called a _ bilateral symmetric interaction ( bsi ) game _ if there exist functions @xmath190 and @xmath191 such that @xmath192 where @xmath193 for every @xmath194 @xcite . a bsi game @xmath195 is an epg with potential function
] @xmath196 consider a _ quasi - cournot game _
@xmath197 with a linear inverse demand function , where each player @xmath38 produces a homogeneous product and determines the output .
let @xmath198 be a set of possible outputs .
the payoff function of player @xmath14 is defined by @xmath199 where @xmath200 and @xmath201 is a differentiable cost function . since @xmath202 @xmath203 is an epg with potential @xmath204 @xcite .
further discussion can be found in @xcite .
[ th : interactionpotential ] a normal form game @xmath205 is an epg if and only if there exists a function @xmath206 ( called an _ interaction potential _ ) such that @xmath207 for every @xmath208 and @xmath77 .
the potential function is @xmath209 in congestion games ( cgs ) , the payoff for using a resource ( e.g. , a channel or a facility ) is a function of the number of players using the same resource .
more precisely , cgs are defined as follows : in the congestion model proposed by rosenthal @xcite , each player @xmath15 uses a subset @xmath3 of common resources @xmath210 , and receives resource - specific payoff @xmath211 from resource @xmath212 according to the number of players using resource @xmath213 . here , @xmath214
, @xmath215 represents the set of players that use resource @xmath213 .
then , @xmath216 . a strategic form game @xmath217 associated with a congestion model , where @xmath218 and @xmath219 is called a cg .
note that @xmath0 is a collection of subsets of @xmath220 and is not a set .
moreover , @xmath1 is a set , not a scalar quantity .
note that a cg where the strategy of every player is a singleton , i.e. , @xmath221 and @xmath222 is called a _
singleton cg_. [ th : congestion ] a cg @xmath223 is an epg with potential function @xmath224 ( * ? ? ?
* theorem 3.1)@xcite .
furthermore , every epg with finite strategy sets has an equivalent cg ( * ? ? ?
* theorem 3.2 ) .
note that generalized cgs do not necessarily possess potential functions . for generalized cgs with potential ,
we refer the interested reader to @xcite .
it was proved that cgs with player - specific payoff functions @xcite , and those with resource - specific payoff functions and player - specific constants @xcite , have potential .
cgs with linear payoff function on undirected / directed graphs has been discussed in @xcite . in some scenarios ,
we can design payoff functions and assign them to players to ensure that the game is an epg .
such approach is often applied in the context of cooperative control @xcite .
these design methodologies can be used when we want to derive payoff functions from a given global objective so that the game with the designed payoff functions is an epg with the global objective as the potential function .
if the global objective is in the form of ( [ eq : interaction_potential ] ) , we can derive payoff functions by using ( [ eq : interaction_payoff_utility ] ) .
otherwise , we can utilize many design rules : the equally shared rule , marginal contribution , and the shapley values @xcite .
since marden and wierman @xcite have already summarized these rules , we only present marginal contribution here . marginal contribution , or the _ wonderful life utility ( wlu ) _
@xcite , is the following payoff function derived from the potential function : @xmath225 where @xmath226 is the value of the potential function in the absence of player @xmath14 .
the game with the wlu is an epg with potential function @xmath227 @xcite .
when the potential function for each player is represented as the sum of functions @xmath228 , i.e. , @xmath229 and @xmath230 , the wlu ( [ eq : wlu1 ] ) can be written as @xmath231 where @xmath232 represents the loss to player @xmath20 resulting from player @xmath14 s participation . in the consensus problem @xcite ,
each player @xmath15 adjusts @xmath3 and tries to reach @xmath233 .
marden et al .
@xcite considered the global objective @xmath234 and proposed using the wlu @xmath235 since game @xmath236 is a bsi game with @xmath237 , @xmath238 is confirmed to be an epg .
in contrast to epgs , opgs have many ordinal potential functions @xcite .
[ eq : ordinal_transform ] consider the game @xmath150 . if there exists a strictly increasing transformation @xmath239 for every @xmath38 such that game @xmath240 is an opg , the original game @xmath150 is an opg with the same potential function @xcite
a variety of learning algorithms are available to facilitate the convergence of potential games to nash equilibrium , e.g. , myopic best response , fictitious play , reinforcement learning , and spatial adaptive play
. unfortunately , there are no general dynamics that are guaranteed to converge to a nash equilibrium for a wide class of games @xcite . since lasaulce et al .
* sections 5 and 6 ) comprehensively summarized these learning algorithms and their sufficient conditions for convergence for various classes of games ( including potential games ) , we present only two frequently used algorithms . _ best - response dynamics
_ refers to the following update rule : at each step @xmath149 , player @xmath77 _ unilaterally _ changes his / her strategy from @xmath145 $ ] to his / her best response @xmath146 $ ] ; in particular , @xmath241 & \in \br_i({\boldsymbol a}_{-i}[k]).\end{aligned}\ ] ] the other players choose the same strategy , i.e. , @xmath242 = { \boldsymbol a}_{-i}[k]$ ] .
note that while the term `` best - response dynamics '' was introduced by matsui @xcite , it has many representations depending on the type of game .
we also note that best - response dynamics may converge to sub - optimal nash equilibria .
by contrast , the following spatial adaptive play can converge to the optimal nash equilibrium . to be precise
, it maximizes the potential function with arbitrarily high probability .
consider a game with a finite number of strategy sets .
_ log - linear learning _
@xcite , _ spatial adaptive play _
@xcite , and _ logit - response dynamics _
@xcite refer to the following update rule : at each step @xmath149 , a player @xmath77 unilaterally changes his / her strategy from @xmath145 $ ] to @xmath243 with probability @xmath244 according to the boltzmann - gibbs distribution @xmath245 ) & = \frac{\exp[{\beta u_i(a_i,{\boldsymbol a}_{-i}[k])}]}{\sum_{a_i ' \in \mathcal{a}_i}\exp{[\beta u_i(a_i',{\boldsymbol a}_{-i}[k ] ) } ] } , \label{eq : logit}\end{aligned}\ ] ] where @xmath246 is related to the ( inverse ) temperature in an analogy to statistical physics .
note that in the limit @xmath247 , the spatial adaptive play approaches the best - response dynamics .
note that ( [ eq : logit ] ) is the solution to the following approximated maximization problem : @xmath248 , \label{eq : perturbed}\end{aligned}\ ] ] which is called a _
payoff , where @xmath249 is the entropy function .
the derivation of ( [ eq : logit ] ) from ( [ eq : perturbed ] ) can be found in @xcite .
[ @xcite ] in the finite epg @xmath250 with potential function @xmath42 , the spatial adaptive play has the unique stationary distribution of strategy profile @xmath251 , where @xmath252}{\sum_{{\boldsymbol a } \in \mathcal{a}}\exp{[\beta \phi({\boldsymbol a})}]},\end{aligned}\ ] ] i.e. , it is also the boltzmann - gibbs distribution
. further discussion can be found in @xcite .
in the tx - rx pair model shown in fig . [
fig : system_models ] , nie and comaniciu @xcite pointed out that the channel selection game @xmath129 introduced in section [ sec : game ] was not an epg .
note that the payoff function of @xmath253 is the negated sum of received interference from neighboring txs . to ensure that the channel selection game is an epg , they considered the channel selection game @xmath254 , whose payoff function was the negated sum of the received interference from neighboring txs , and generated interference to neighboring rxs , i.e. , @xmath255
since @xmath256 is a bsi game with @xmath257 , it is an epg with potential @xmath258 which corresponds to the negated sum of received interference in the entire network .
note that in order to evaluate ( [ eq : nie_utility ] ) , each pair @xmath14 needs to estimate or share the values of the generated interference to neighboring rxs , @xmath259 .
concurrently with the above , kauffmann et al .
@xcite discussed the following potential function @xmath260 , which includes rx - specific noise power @xmath261 , and derived a payoff function using theorem [ th : interactionpotential ] , @xmath262 to enable multi - channel allocation , e.g. , orthogonal frequency - division multiple access ( ofdma ) subcarrier allocation or resource block allocation , la et al .
@xcite discussed a modification of @xmath263 suitable for multi - channel allocation .
in contrast to unidirectional links assumed in the tx - rx pair model , uykan and jntti @xcite discussed a channel assignment problem for bidirectional links and proposed a joint transmission order and channel assignment algorithm .
nie et al .
@xcite showed that the joint channel selection and power control game with payoff function @xmath264 is an epg . because the best response in @xmath265 results in the minimum transmission power level , bloem et al .
@xcite proposed adding terms @xmath266 to ( [ eq : nie2_utility ] ) to account for the achievable data rate and consumed power .
note that these terms are self - motivated functions , and the game with the modified payoff function is still an epg .
as another type of joint assignment , a preliminary beamform pattern setting followed by channel allocation was discussed in @xcite . to manage interference in primary - secondary systems , bloem et al .
@xcite proposed adding terms related to the received and generated interferences from and to the primary user .
they also proposed adding cost terms related to payoff function ( [ eq : nie2_utility ] ) .
in particular , they discussed a stackelberg game @xcite , where the primary user was the leader and the secondary users were followers .
giupponi and ibars discussed overlay cognitive networks @xcite and heterogeneous ofdma networks @xcite .
mustika et al .
@xcite took a similar approach to prioritize users .
uplinks of heterogeneous ofdma cellular systems with femtocells were discussed in @xcite , whereas downlinks of ofdma cellular systems , where each bs transmits to several mobile stations , were discussed in @xcite .
ofdma relay networks were considered in @xcite .
further discussion can be found in @xcite .
joint bs / ap selection and channel selection problems were discussed in @xcite .
in the tx - rx pair model shown in fig . [
fig : system_models ] , the signal - to - interference - plus - noise ratio ( sinr ) at rx @xmath15 is given by @xmath267 menon et al .
@xcite pointed out that there may be no nash equilibrium in the channel selection game @xmath268 .
instead , they proposed using the sum of the inverse sinr , defined by @xmath269 as the payoff function .
similar to @xmath270 , @xmath271 is a bsi game with @xmath272 \indicator { c_j = c_i } $ ] .
thus , @xmath273 is an epg with potential @xmath274 i.e. , the sum of the inverse sinr in the network .
note that the above expression is a single carrier version of orthogonal channel selection .
menon et al .
@xcite discussed a waveform adaptation version of @xmath275 that can be applied to codeword selection in non - orthogonal code division multiple access ( cdma ) , and buzzi et al . @xcite
further discussed waveform adaptation .
buzzi et al .
@xcite also discussed an ofdma subcarrier allocation version of @xmath276 .
cai et al .
@xcite discussed joint transmission power and channel assignment utilizing the payoff function ( [ eq : buzzi_utility ] ) of @xmath276 .
gllego et al .
@xcite proposed using the network throughput of joint power and channel assignment , @xmath277 as potential , where @xmath278 is the bandwidth of channel @xmath279 , and @xmath73 is the required sinr .
it may have been difficult to derive a simple payoff function , and they thus proposed the wlu ( [ eq : wlu2 ] ) of ( [ eq : gallego ] ) .
yu et al . @xcite and chen et al .
@xcite considered sensor networks where each rx ( sink ) receives messages from multiple txs ( sensors ) .
they proved that a channel selection that minimizes the number of received and generated interference signals is an epg , where the potential is the number of total interference signals .
note that the average number of retries is approximately proportional to the number of received interference signals when the probability that the messages are transmitted is very small , as in sensor networks .
a simpler and related form of ( [ eq : nie_utility ] ) is detailed in the following discussion . to reduce the information exchange required to evaluate ( [ eq : nie_utility ] ) , yamamoto et al .
@xcite proposed using the number of received and generated interference sources as the payoff function , where the received interference power is greater than a given threshold @xmath280 , i.e. , @xmath281 this model is sometimes referred to as a `` binary '' interference model @xcite in comparison with a `` physical '' interference model . because @xmath282 is a bsi game with @xmath283 , @xmath284 is an epg .
when we consider a directed graph , where edges between tx @xmath20 and rx @xmath15 indicate @xmath285 , we denote tx @xmath15 s neighboring rxs by @xmath286 , and rx @xmath15 s neighboring txs by @xmath287 . using these expressions , ( [ eq : yamamoto_utility ] ) can be rewritten to @xmath288 yang et al .
@xcite discussed a multi - channel version of @xmath289 .
in section [ ssec : nie ] , channel allocation games in the tx - rx pair model shown in fig . [
fig : system_models ] are discussed .
neel et al .
@xcite considered a different channel allocation game typically applied to channel allocation for aps in the wireless local area networks ( wlans ) shown in fig .
[ fig : system_models ] , where each tx @xmath77 selects a channel @xmath290 to minimize the interference from other txs , i.e. , @xmath291 where @xmath292 is the common transmission power level for every tx .
note that @xmath18 in this scenario , whereas @xmath293 in the tx - rx pair model shown in fig .
[ fig : system_models ] .
moreover , note that interference from stations other than the txs is not taken into account in the payoff function .
in addition to the tx network model , channel selection can be applied to the canonical network model shown in fig .
[ fig : system_models ] @xcite . because @xmath294 is a bsi game where @xmath295
, it is an epg with potential @xmath296 which corresponds to the aggregated interference power among txs .
neel et al .
pointed out that other symmetric interference functions , e.g. , @xmath297 , where @xmath298 is the common bandwidth for every channel , can be used instead of @xmath299 in ( [ eq : neel_utility ] ) .
kauffmann et al .
@xcite discussed essentially the same problem .
however , they considered player - specific noise , and derived ( [ eq : neel_utility ] ) by substituting @xmath18 and @xmath300 into ( [ eq : kauffmann_utility ] ) .
compared with the payoff function ( [ eq : nie_utility ] ) , ( [ eq : neel_utility ] ) can be evaluated with only local information available at each tx ; however , the transmission power levels of all txs need to be identical .
we further discuss this requirement in section [ ssec : non - identical ] .
liu and wu @xcite reformulated the game represented by ( [ eq : neel_utility ] ) as a cg by introducing virtual resources .
further discussion can be found in @xcite . to avoid the requirement of identical transmission power levels in ( [ eq : neel_utility ] )
, neel @xcite proposed using the product of ( constant ) transmission power level @xmath72 and interference @xmath301 as the payoff function , i.e. , @xmath302 because @xmath303 is a bsi game with @xmath304 , @xmath305 is an epg with @xmath306 note that this form of payoff functions was provided by menon et al .
@xcite in the context of waveform adaptations .
this game under frequency - selective channels was discussed by wu et al .
@xcite .
the relationship between ( [ eq : babadi_utility ] ) and its exact potential function ( [ eq : babadi_v ] ) implies that the game @xmath307 with payoff function @xmath308 is a wpg with potential function @xmath309 and @xmath310 in ( [ eq : wpg ] ) , i.e. , the identical transmission power level required in ( [ eq : neel_utility ] ) is not necessarily required for the game to have the fip .
this was made clear by bahramian et al .
@xcite and babadi et al .
@xcite . as extensions , in @xcite , the interference management game @xmath311 on graph structures with the following payoff function
was discussed : @xmath312 @xcite proposed using the expected value of interference in order to manage fluctuating interference .
zheng @xcite treated dynamical on - off according to traffic variations in @xmath313 .
menon et al . @xcite showed that a waveform adaptation game where the payoff function is the sinr or the mean - squared error at the rx is an opg .
chen and huang @xcite showed that a channel allocation game in the tx network model shown in fig .
[ fig : system_models ] , or in the canonical network model shown in fig .
[ fig : system_models ] , where the payoff function is the sinr or a shannon capacity , is an opg . here
, we provide a derivation in the form of channel allocation according to the derivation provided in @xcite .
a channel selection game @xmath314 with payoff function @xmath315 \label{eq : pin_utility}\end{aligned}\ ] ] is an epg with potential @xmath316 because @xmath317 is a constant in ( [ eq : pin_utility ] ) , by theorem [ eq : ordinal_transform ] , @xmath318 with payoff @xmath319 is an opg with potential @xmath320 . as a result , once again using theorem [ eq : ordinal_transform ] , @xmath321 with payoff @xmath322 is an opg with potential @xmath320 .
@xcite further discuss @xmath323 , where the active tx set can be stochastically changed .
a quite relevant discussion was conducted by song et al .
they discussed a joint transmission power and channel assignment game @xmath324 to maximize throughput : @xmath325 where @xmath326 represents throughput depending on sinr .
they pointed out that since each user would set the maximum transmission power at a nash equilibrium , @xmath327 is equivalent to the channel selection game @xmath328 .
further discussion on joint transmission power and channel assignment can be found in @xcite .
for the interference graph @xmath21 shown in fig .
[ fig : system_models ] , xu et al .
@xcite proposed using the number of neighbors that select the same channel as the payoff function , i.e. , @xmath329 we would like to point out that ( [ eq : xustsp1_utility ] ) can be reformulated to @xmath330 i.e. , @xmath331 is a bsi game with @xmath332 . thus @xmath333 is an epg .
note that this is a special case of singleton cgs on graphs discussed in section [ ssec : aloha ] . as variations of @xmath333 , xu et al .
@xcite discussed the impact of partially overlapped channels .
yuan et al .
@xcite discussed the variable - bandwidth channel allocation problem .
zheng et al .
@xcite took into account stochastic channel access according to the carrier sense multiple access ( csma ) protocol .
@xcite discussed a multi - channel version of @xmath334 .
liu et al .
@xcite discussed a common control channel assignment problem for cognitive radios , and proposed using @xmath335 for the payoff function so that every player chooses the same channel .
this game is similar to the consensus game @xmath238 .
channels can be viewed as common resources in the congestion model introduced in section [ sssec : congestion_game ] . in general , throughput when using a channel depends only on the number of stations that select the relevant channel .
a cg formulation is thus frequently used for channel selection problems .
altman et al .
@xcite formulated a multi - channel selection game in a single collision domain as a cg . based on a cg formulation , channel selections by secondary stations were discussed in @xcite . a channel selection problem in multiple collision domains
was discussed in @xcite .
iellamo et al .
@xcite used numerically evaluated successful access probabilities depending on the number of stations in csma / ca as payoff functions . here , we discuss channel selection problems in interference graph @xmath21 , where each node @xmath38 attempts to adjust its channel @xmath279 to maximize its successful access probability or throughput . consider collision channels shared using slotted aloha .
each node @xmath15 adjusts its channel to avoid simultaneous transmissions on the same channel because these result in collisions . in this case , when one node exclusively chooses a channel , the node can transmit without collisions .
thus , the following payoff function captures the benefit of nodes : @xmath336 @xmath337 is a singleton cg on graphs , and thomas et al .
@xcite showed that @xmath338 is an opg is equivalent to @xmath339 when setting @xmath340 for every @xmath15 . ] .
consider that each node has a transmission probability @xmath341 ( @xmath342 ) .
chen and huang @xcite proposed using the logarithm of successful access probability , @xmath343\end{aligned}\ ] ] and proved that @xmath344 is a wpg . here , we provide a different proof .
when we consider @xmath345 \nonumber \\ & = - \log ( 1 - x_i ) \log ( x_i ) \nonumber\\ & \hphantom { = { } } - \log ( 1 - x_i ) \textstyle\sum _ { j \neq i } \indicator { c_j = c_i } \indicator { ij \in \mathcal { e } } \log ( 1 - x_j ) , \nonumber\end{aligned}\ ] ] @xmath346 is a bsi game with @xmath347 .
thus , @xmath348 is a wpg and , by theorem [ eq : ordinal_transform ] , @xmath349 with payoff @xmath350 is an opg . chen and huang @xcite further discussed @xmath339 with player - specific constants and proved that the game is an opg . before concluding this section
, we would like to point out the relationship between @xmath351 and cgs .
when we assume an identical transmission probability @xmath352 for every @xmath15 , we get @xmath353 i.e. , @xmath354 is a cg on graphs .
let the backoff time of player @xmath15 be denoted by @xmath355 $ ] , where @xmath356 represents the backoff window size .
the probability to acquire channel access is given by @xmath357 @xmath358 is a singleton cg , and thus is an epg .
furthermore , @xmath359 with @xmath360 is also a singleton cg .
chen and huang @xcite showed that @xmath361 with player - specific constants is an opg .
they @xcite further discussed @xmath361 and @xmath362 on graphs with player - specific constants , and proved that these are opgs according to the proof provided in @xcite .
@xcite further discussed the game under fading channels .
chen and huang @xcite generalized @xmath362 to @xmath363 , whose payoff function is a generalized throughput @xmath364 where @xmath365 represents the channel - sharing weight for player @xmath15 .
du et al . @xcite
further discussed this kind of game . for the tx - rx pair model shown in fig .
[ fig : system_models ] , canales and gllego @xcite proposed using the following network throughput as a result of joint transmission power and channel assignment as potential : @xmath366 where @xmath16 , @xmath278 is the bandwidth of channel @xmath279 , and @xmath26 is the power threshold of interference .
since ( [ eq : canales ] ) is too complex , it may be difficult to derive simple payoff functions .
thus , canales and gllego proposed using payoff functions of the form of a wlu ( [ eq : wlu2 ] ) .
note that the evaluation of the wlu of ( [ eq : canales ] ) requires the impact of joint assignment on the throughput of neighboring nodes .
consider a collision channel shared using slotted aloha .
each node @xmath15 adjusts its transmission probability @xmath367 $ ] to maximize the following successful access probability ( minus the cost ) : @xmath368 this is a well - known payoff function .
further discussion can be found in @xcite . because ( [ eq : saloha_utility ] ) satisfies ( [ eq : twice_difference ] ) , @xmath369 ) , ( u{{\ref{eq : saloha}}}_i ) ) $ ] is an epg with potential @xmath370 candogan et al .
@xcite showed that @xmath371 in stochastic channel model , where each player adjusts his / her transmission probability based on the channel state , is a wpg .
cohen et al .
@xcite discussed a multi - channel version of @xmath372 .
they also discussed @xmath373 on graphs @xcite . for this kind of transmission probability adjustment to satisfy @xmath374 , the cost function
@xmath375 needs to be used @xcite . because this cost function is a coordination function , a game with this cost function is still an epg .
here , we discuss power control problems in multiple - access channels , as shown in fig . [
fig : system_models ] , where each tx attempts to adjust its transmission power level to maximize its throughput . for a summary of transmission power control ,
we refer to @xcite .
note that saraydar et al .
@xcite applied a game - theoretic approach to an uplink transmission power control problem in a cdma system .
the relation between potential games and transmission power control to achieve target sinr or target throughput has been discussed in @xcite .
alpcan et al .
@xcite formulated uplink transmission power control in a single - cell cdma as the game @xmath376 , where @xmath377 , @xmath378 is the minimum transmission power , and @xmath379 is the maximum transmission power . in this game , each tx @xmath38 adjusts its transmission power @xmath380 to maximize its data rate ( throughput ) , which is assumed to be proportional to the shannon capacity , minus the cost of transmission power , i.e. , @xmath381 where @xmath382 is the spreading gain and @xmath383 is a positive real number . the cost function @xmath384 is used to avoid an inefficient nash equilibrium , where all txs choose the maximum transmission power .
all txs choose this power because @xmath385 is the maximum transmission power for every tx when the cost function is not used @xcite .
alpcan et al .
@xcite proved the existence and uniqueness of a nash equilibrium in the game , and neel @xcite showed that this game is not an epg because ( [ eq : twice_difference ] ) does not hold .
note that neel et al .
@xcite was the first to apply the potential game approach to this type of power control . instead of @xmath386 , fattahi and
paganini @xcite proposed setting @xmath387 in @xmath386 , i.e. , @xmath388 where @xmath389 is a non - decreasing convex cost function .
since @xmath390 is a linear combination of a coordination function @xmath391 , a dummy function @xmath392 , and a self - motivated function @xmath393 , @xmath394 is an epg with potential @xmath395 because @xmath396 is continuously differentiable and strictly concave , by theorem [ th : unique ] , there is a unique maximizer for the potential , and best - response dynamics converge to a unique nash equilibrium , which is the maximizer of the potential on strategy space @xmath397 .
kenan et al .
@xcite discussed @xmath398 over time - varying channels .
neel @xcite approximated ( [ eq : cdma_utility ] ) by @xmath399 and showed that @xmath400 is an epg with potential @xmath401 .
candogan et al .
@xcite applied @xmath402 to multi - cell cdma systems , and verified that the modified game is an epg with a unique nash equilibrium by applying theorem [ th : unique ] .
a more general form of payoff functions of sinr was discussed in @xcite .
a transmission power control game @xmath403 with multiple channels was discussed in @xcite .
let the set of channels be denoted by @xmath404 .
each tx @xmath77 transmits through a subset of @xmath405 to maximize the aggregated capacity @xmath406 by adjusting the transmission power vector @xmath407 .
this game is an epg with potential @xmath408 .
mertikopoulos et al .
@xcite further discussed the game under fading channels .
note that multi - channel transmission power assignment problems can be seen as joint transmission power and channel assignment problems introduced in section [ eq : neel_capacity ] because a zero transmission power level means that the relevant channel has not been assigned @xcite .
note that @xcite also discussed bs selection , and further discussion can be found in @xcite .
the joint transmission power and bandwidth assignment problem for relay networks was discussed in @xcite .
primary - secondary scenario @xcite and heterogeneous network scenario @xcite were also discussed .
closely related problems to the power control problems discussed above are found in precoding schemes for multiple - input multiple - output ( mimo ) multiple - access channels .
the instantaneous mutual information of tx @xmath15 , assuming that multiuser interference can be modeled as a gaussian random variable , is expressed as @xmath409 where @xmath410 , @xmath411 is the channel matrix , @xmath412 is the hermitian transpose of @xmath413 , @xmath414 is a covariance matrix of input signal , @xmath415 is the number of antennas at every tx , and @xmath416 is the number of antennas at a single rx .
belmega et al .
@xcite discussed a game where an input covariance matrix @xmath414 is adjusted .
concurrently , zhong et al .
@xcite discussed a game where a precoding matrix is adjusted .
since @xmath414 is calculated from a precoding matrix , these games are equivalent .
since ( [ eq : mutual_inf ] ) is a coordination - dummy function , this game is an epg with the system s achievable sum - rate as potential .
this game was further discussed in ( * ? ? ?
* section 8) .
energy efficiency @xcite , primary - secondary scenario @xcite , and relay selection @xcite were also discussed .
joint precoding and ap selection in multi - carrier system was discussed in @xcite .
the primary goal of topology control @xcite is to adjust transmission power to maintain network connectivity while reducing energy consumption to extend network lifetime and/or reducing interference to enhance throughput .
komali et al .
@xcite formulated the topology control problem in the tx network model shown in fig .
[ fig : system_models ] as @xmath417 with @xmath418 $ ] and @xmath419 where @xmath420 , and @xmath421 is the number of txs with whom tx @xmath15 establishes ( possibly over multiple hops ) a communication path using bidirectional links . note that @xmath422 when @xmath423 .
this game has been shown to be an opg with @xmath424 note that the mathematical representation of @xmath425 using connectivity matrix @xcite was first proposed in @xcite .
komali et al .
@xcite also discussed interference reduction through channel assignment , which is seen as a combination of @xmath426 and a channel assignment version of @xmath427 .
they further discussed the impact of the amount of knowledge regarding the network on the spectral efficiency @xcite .
chu and sethu @xcite considered battery - operated stations and formulated transmission power control to prolong network lifetime while maintaining connectivity as an opg .
similar approaches can be found in @xcite , and the joint assignment of transmission power and channels was discussed in @xcite .
liu et al .
@xcite formulated measures for transmission power and sensing range adjustment to enhance energy efficiency while maintaining sensor coverage as an opg .
baar et al .
@xcite formulated a flow and congestion control game , where each user @xmath15 adjusts the amount of traffic flow @xmath428 to enhance @xmath429 where @xmath430 represents the commodity - link cost of congestion .
because ( [ eq : basar_utility ] ) is a combination of self - motivated and coordination functions , a game @xmath431 with payoff function @xmath432 is an epg @xcite with @xmath433 the learning process of this game was further discussed by scutari et al .
other payoff functions for flow control were discussed in @xcite .
douligeris and mazumdar @xcite , and zhang and douligeris @xcite introduced an m / m/1 queuing game @xmath434 , where each user @xmath14 transmits packets to a single server at departure rate @xmath435 and adjusts the arrival rate @xmath436 to maximize the `` power '' @xcite , which is defined as the throughput @xmath436 divided by the delay @xmath437 , i.e. , @xmath438 where @xmath439 is a factor that controls the trade - off between throughput and delay .
note that this game is a cournot game ( see ( [ eq : cournot ] ) ) when @xmath440 for every @xmath15 .
gai et al .
@xcite proved that @xmath441 is an opg . here , we provide a different proof . because a game with payoff function @xmath442 is an epg , by theorem [ eq : ordinal_transform ]
, @xmath443 is an opg .
marden et al . @xcite pointed out that the sensor deployment problem ( see @xcite and references therein ) , where each mobile node @xmath14 updates its location @xmath444 to forward data from immobile sources to immobile destinations , can be formulated as an epg .
since the required transmission power to an adjacent node @xmath445 is proportional to the square of the propagation distance , @xmath446 , in a free - space propagation environment , minimizing the total required transmission power problem is formulated as a maximization problem with global objective @xmath447 if @xmath448 is used as the payoff function of node @xmath15 , @xmath449 is equivalent to the consensus game @xmath238 .
a sensor coverage problem is formulated as a maximization problem with global objective in continuous form @xmath450
@xcite @xmath451 \ , \dd r,\end{aligned}\ ] ] or in discrete form @xcite @xmath452,\end{aligned}\ ] ] where @xmath453 is the specific region to be monitored , @xmath454 is an event density function or value function that indicates the probability density of an event occurring at point @xmath455 , @xmath456 $ ] is the probability of sensor @xmath15 to detect an event occurring at @xmath455 , and @xmath457 is the location of sensor @xmath15 . for a summary of coverage problems ,
we refer the interested reader to @xcite .
arslan et al .
@xcite discussed a game where each mobile sensor @xmath15 updates its location @xmath457 , treated @xmath458 as potential , and proposed assigning a wlu to each sensor , i.e. , @xmath459 where @xmath460 corresponds to the probability that sensor @xmath15 detects an event occurring at @xmath461 alone .
further discussion can be found in @xcite .
we would like to note that @xmath462 has a similar expression with @xmath463 . in the same manner in @xmath464 , drr et al .
@xcite treated @xmath465 as potential and proposed assigning a wlu @xmath466\ , \dd r. \label{eq : coverage_continuous_utility}\end{aligned}\ ] ] zhu and martnez @xcite considered mobile sensors with a directional sensing area .
each mobile sensor updates its location and direction .
the reward from a target is fairly allocated to sensors covering the target .
arsie et al .
@xcite considered a game where each node @xmath15 attempts to maximize the expected value of the reward . here , each node @xmath15 receives the reward if node @xmath15 is the first to reach point @xmath461 , and the value of the reward is the time until the second node arrives , i.e. , @xmath467 @xmath468 was proved to be an epg .
@xcite formulated a time slot assignment problem for immobile sensors , which is equivalent to a channel allocation problem , as @xmath469 , where each sensor @xmath38 selects a slot @xmath470 to maximize the area covered only by sensor @xmath15 , i.e. , @xmath471 where @xmath472 is the sensing area covered by sensor @xmath14 .
game @xmath473 was proved to be an epg with potential @xmath474 where @xmath475 corresponds to the average coverage . to show the close relationship between the payoff functions ( [ eq : saloha_utility ] ) in the slotted aloha game @xmath463 and ( [ eq : coverage_continuous_utility ] ) in the coverage game @xmath476
, we provide different expressions . using @xmath477 , we get @xmath478 where the surface integral is taken over the whole area .
wang et al . @xcite further discussed this problem .
song et al .
@xcite applied the coverage game to camera networks .
ding et al .
@xcite discussed a pan - tilt - zoom ( ptz ) camera network to track multiple targets .
another potential game - theoretic ptz camera control scheme was proposed in @xcite , motivated by natural environmental monitoring .
directional sensors were discussed in @xcite .
the form of payoff functions is similar to ( [ eq : coverage_discrete_utility ] ) . until now
, each immobile sensor was assumed to receive a payoff when it covered a target alone .
yen et al .
@xcite discussed a game where each sensor receives a payoff when the number of sensors covering a target is smaller than or equal to the allowable number .
since this game falls within a class of cgs , it is also an epg .
we have provided a comprehensive survey of potential game approaches to wireless networks , including channel assignment problems and transmission power assignment problems .
although there are a variety of payoff functions that have been proven to have potential , there are some representative forms , e.g. , bsi games and congestion games , and we have shown the relations between representative forms and individual payoff functions .
we hope the relations shown in this paper will provide insights useful in designing wireless technologies .
other problems that have been formulated in terms of potential games are found in routing @xcite , bs / ap selection @xcite , cooperative transmissions @xcite , secrecy rate maximization @xcite , code design for radar @xcite , broadcasting @xcite , spectrum market @xcite , network coding @xcite , data cashing @xcite , social networks @xcite , computation offloading @xcite , localization @xcite , and demand - side management in smart grids @xcite .
this work was supported in part by jsps kakenhi grant numbers 24360149 , 15k06062 .
the author would like to acknowledge dr .
takeshi hatanaka at tokyo institute of technology for his insightful comments on cooperative control and ptz camera control .
the author acknowledges dr .
i. wayan mustika at universitas gadjah mada , and dr .
masahiro morikura and dr .
takayuki nishio at kyoto university for their comments .
h. al - tous and i. barhumi , `` joint power and bandwidth allocation for amplify - and - forward cooperative communications using stackelberg game , '' _ ieee trans .
_ , vol .
62 , no . 4 , pp .
16781691 , may 2013 .
t. alpcan and t. baar , `` a game - theoretic framework for congestion control in general topology networks , '' in _ proc .
41st ieee conf . decision and control ( cdc )
_ , vol . 2 , las vegas , nv , dec . 2002 , pp . 12181224 .
e. altman , y. hayel , and h. kameda , `` evolutionary dynamics and potential games in non - cooperative routing , '' in _ proc .
modeling optim .
mobile , ad hoc , wireless netw .
( wiopt ) _ , limassol , apr .
2007 , pp . 15 .
e. altman , a. kumar , and y. hayel , `` a potential game approach for uplink resource allocation in a multichannel wireless access network , '' in _ proc .
icst conf .
tools ( valuetools ) _ , pisa , italy , oct . 2009 , pp .
72:172:9 .
g. arslan , m. f. demirkol , and y. song , `` equilibrium efficiency improvement in mimo interference systems : a decentralized stream control approach , '' _ ieee trans .
wireless commun .
_ , vol . 6 , no . 8 , pp . 29842993 , aug .
2007 .
t. baar and r. srikant , `` revenue - maximizing pricing and capacity expansion in a many - users regime , '' in _ proc .
( infocom ) _ , vol . 1 , new york , ny , jun .
2002 , pp . 294301 .
m. bennis , s. m. perlaza , p. blasco , z. han , and h. v. poor , `` self - organization in small cell networks : a reinforcement learning approach , '' _ ieee trans .
wireless commun .
_ , vol . 12 , no . 7 , pp . 32023212 , jul
2013 .
m. bloem , t. alpcan , and t. baar , `` a stackelberg game for power control and channel allocation in cognitive radio networks , '' in _ proc .
workshop game theory commun .
( gamecomm ) _ , nantes , france , oct .
2007 , pp . 19 .
s. buzzi , g. colavolpe , d. saturnino , and a. zappone , `` potential games for energy - efficient power control and subcarrier allocation in uplink multicell ofdma systems , '' _ ieee j. sel .
topics signal process . _ ,
vol . 6 , no . 2 , pp
. 89103 , apr .
2012 .
y. cai , j. zheng , y. wei , y. xu , and a. anpalagan , `` a joint game - theoretic interference coordination approach in uplink multi - cell ofdma networks , '' _ wireless pers .
_ , vol .
80 , no . 3 , pp .
12031215 , feb .
2015 .
m. canales , j. ortn , and j. r. gllego , `` game theoretic approach for end - to - end resource allocation in multihop cognitive radio networks , '' _ ieee commun .
_ , vol . 16 , no . 5 , pp .
654657 , may 2012 .
u. o. candogan , i. menache , a. ozdaglar , and p. a. parrilo , `` competitive scheduling in wireless collision channels with correlated channel state , '' in _ proc .
game theory netw .
( gamenets ) _ , istanbul , may 2009 , pp . 621630 .
j. chen , q. yu , p. cheng , y. sun , y. fan , and x. shen , `` game theoretical approach for channel allocation in wireless sensor and actuator networks , '' _ ieee trans .
56 , no .
23322344 , oct . 2011 .
x. chen , x. gong , l. yang , and j. zhang , `` a social group utility maximization framework with applications in database assisted spectrum access , '' in _ proc .
( infocom ) _ , apr .
2014 , pp . 19591967 .
x. chu and h. sethu , `` cooperative topology control with adaptation for improved lifetime in wireless ad hoc networks , '' in _ proc .
comput . commun .
( infocom ) _ , orlando , fl , mar .
2012 , pp . 262270 .
k. cohen , a. nedi , and r. srikant , `` distributed learning algorithms for spectrum sharing in spatial random access networks , '' in _ proc .
modeling optim .
mobile , ad hoc , wireless netw .
( wiopt ) _ , may 2015 .
h. dai , y. huang , and l. yang , `` game theoretic max - logit learning approaches for joint base station selection and resource allocation in heterogeneous networks , '' _ to appear in ieee j. sel .
areas commun .
_ , 2015 .
e. del re , g. gorni , l. s. ronga , and r. suffritti , `` resource allocation in cognitive radio networks : a comparison between game theory based and heuristic approaches , '' _ wireless pers .
_ , vol .
49 , no . 3 , pp .
375390 , may 2009 .
m. derakhshani and t. le - ngoc , `` distributed learning - based spectrum allocation with noisy observations in cognitive radio networks , '' _ ieee trans .
_ , vol . 63 , no . 8 , pp .
37153725 , oct .
2014 .
b. f. duarte , z. m. fadlullah , a. v. vasilakos , and n. kato , `` on the partially overlapped channel assignment on wireless mesh network backbone : a game theoretic approach , '' _ ieee j. sel .
areas commun . _ ,
30 , no . 1 ,
119127 , jan . 2012 .
drr , m. s. stankovi , and k. h. johansson , `` distributed positioning of autonomous mobile sensors with application to coverage control , '' in _ proc .
control conf .
( acc ) _ , san francisco , ca , jun .
2011 , pp . 48224827 .
y. gai , h. liu , and b. krishnamachari , `` a packet dropping - based incentive mechanism for m / m/1 queues with selfish users , '' in _ proc .
( infocom ) _ , shanghai , apr .
2011 , pp . 26872695 .
j. r. gllego , m. canales , and j. ortn , `` distributed resource allocation in cognitive radio networks with a game learning approach to improve aggregate system capacity , '' _ ad hoc netw . _ , vol .
10 , no . 6 , pp . 10761089 , aug . 2012 .
a. ghosh , l. cottatellucci , and e. altman , `` normalized nash equilibrium for power allocation in femto base stations in heterogeneous network , '' in _ proc .
modeling optim .
mobile , ad hoc , wireless netw . ( wiopt )
_ , mumbai india , may 2015 .
hao , y .- x
. zhang , n. jia , and b. liu , `` virtual game - based energy balanced topology control algorithm for wireless sensor networks , '' _ wireless pers .
_ , vol .
69 , no . 4 , pp . 12891308 , apr .
hao , y .- x .
zhang , and b. liu , `` distributed cooperative control algorithm for topology control and channel allocation in multi - radio multi - channel wireless sensor network : from a game perspective , '' _ wireless pers .
_ , vol .
73 , no . 3 , pp .
353379 , dec . 2013 .
t. hatanaka , y. wasa , and m. fujita , `` game theoretic cooperative control of ptz visual sensor networks for environmental change monitoring , '' in _ proc .
52nd ieee conf .
decision and control ( cdc ) _ , firenze , dec .
2013 , pp . 76347640 .
g. he , l. cottatellucci , and m. debbah , `` the waterfilling game - theoretical framework for distributed wireless network information flow , '' _ eurasip j. wirel .
2010 , no . 1 ,
113 , jan . 2010 . h.
he , j. chen , s. deng , and s. li , `` game theoretic analysis of joint channel selection and power allocation in cognitive radio networks , '' in _ proc .
oriented wireless netw .
( crowncom ) _ , may 2008 , pp . 15 .
m. hong , a. garcia , j. barrera , and s. g. wilson , `` joint access point selection and power allocation for uplink wireless networks , '' _ ieee trans . signal process .
_ , vol .
61 , no . 13 , pp . 33343347 , jul .
2013 . c. ibars , m. navarro , and l. giupponi ,
`` distributed demand management in smart grid with a congestion game , '' in _ proc .
. smart grid commun .
( smartgridcomm ) _ , gaithersburg , md , oct .
2010 , pp . 495500 .
l. jiao , x. zhang , o .- c .
granmo , and b. j. oommen , `` a bayesian learning automata - based distributed channel selection scheme for cognitive radio networks , '' in _ proc .
other applications applied intelligent syst . _ , kaohsiung , taiwan , jun .
2014 , pp .
4857 .
b. kauffmann , f. baccelli , a. chaintreau , v. mhatre , k. papagiannaki , and c. diot , `` measurement - based self organization of interfering 802.11 wireless access networks , '' in _ proc .
ieee int . conf .
( infocom ) _ , anchorage , ak , 2007 , pp .
14511459 .
r. s. komali and a. b. mackenzie , `` analyzing selfish topology control in multi - radio multi - channel multi - hop wireless networks , '' in _ proc .
conf . commun .
( icc ) _ , dresden , jun .
2009 , pp .
r. s. komali , r. w. thomas , l. a. dasilva , and a. b. mackenzie , `` the price of ignorance : distributed topology control in cognitive networks , '' _ ieee trans .
wireless commun .
_ , vol .
4 , pp . 14341445 , apr .
o. korcak , t. alpcan , and g. iosifidis , `` collusion of operators in wireless spectrum markets , '' in _ proc .
modeling optim .
mobile , ad hoc , wireless netw .
( wiopt ) _ , paderborn , germany , may 2012 , pp .
3340 . q. d. la , y. h. chew , and b. h. soong , `` an interference - minimization potential game for ofdma - based distributed spectrum sharing systems , '' _ ieee trans .
_ , vol . 60 , no . 7 , pp
. 33743385 , sep .
2011 .
q. d. la , y. h. chew , and b .- h .
soong , `` subcarrier assignment in multi - cell ofdma systems via interference minimization game , '' in _ proc .
ieee wireless commun .
netw . conf .
( wcnc ) _ , shanghai , apr .
2012 , pp . 13211325 .
d. li and j. gross , `` distributed tv spectrum allocation for cognitive cellular network under game theoretical framework , '' in _ proc .
new frontiers dynamic spectrum access netw .
( dyspan ) _ , bellevue , wa , oct .
2012 , pp . 327338 .
j. li , w. liu , and k. yue , `` a game - theoretic approach for balancing the tradeoffs between data availability and query delay in multi - hop cellular networks , '' in _ proc .
theory and appl .
models of comput .
( tamc ) _ , beijing , china , may 2012 , pp .
487497 .
j. li , k. yue , w. liu , and q. liu , `` game - theoretic based distributed scheduling algorithms for minimum coverage breach in directional sensor networks , '' _ int .
j. distrib .
sens . n. _ , vol .
2014 , pp .
112 , may 2014 .
q. liang , x. wang , x. tian , f. wu , and q. zhang , `` two - dimensional route switching in cognitive radio networks : a game - theoretical framework , '' _ to appear in ieee / acm trans .
_ , vol .
pp , no . 99 , 2014 .
d. ling , z. lu , y. ju , x. wen , and w. zheng , `` a multi - cell adaptive resource allocation scheme based on potential game for icic in lte - a , '' _ int . j. commun .
_ , vol . 27 , no . 11 , pp
27442761 , nov .
2014 .
l. liu , r. wang , and f. xiao , `` topology control algorithm for underwater wireless sensor networks using gps - free mobile sensor nodes , '' _ j. netw
_ , vol .
35 , no . 6 , pp .
19531963 , nov .
2012 .
y. liu , l. dong , and r. j. marks ii , `` common control channel assignment in cognitive radio networks using potential game theory , '' in _ proc .
ieee wireless commun .
( wcnc ) _ , shanghai , apr .
2013 , pp . 315320 .
s. maghsudi and s. staczak , `` joint channel selection and power control in infrastructureless wireless networks : a multi - player multi - armed bandit framework , '' _ to appear in ieee trans . veh .
_ , vol .
pp , no . 99 , 2014 .
r. maheshwari , s. jain , and s. r. das , `` a measurement study of interference modeling and scheduling in low - power wireless networks , '' in _ proc .
6th acm conf .
embedded netw .
sensor syst .
( sensys ) _ , raleigh , nc , nov .
2008 , pp . 141154 .
m. mavronicolas , i. milchtaich , b. monien , and k. tiemann , `` congestion games with player - specific constants , '' in _ proc .
math . found .
( mfcs ) _ , czech republic , aug .
2007 , pp . 633644 .
r. menon , a. b. mackenzie , r. m. buehrer , and j. h. reed , `` a game - theoretic framework for interference avoidance in ad hoc networks , '' in _ proc .
ieee global telecommun .
( globecom ) _ , san francisco , ca , nov .
2006 , pp . 16 .
p. mertikopoulos , e. v. belmega , a. l. moustakas , and s. lasaulce , `` distributed learning policies for power allocation in multiple access channels , '' _ ieee j. sel .
areas commun .
_ , vol . 30 , no . 1 , pp .
96106 , jan .
2012 .
n. moshtagh , r. mehra , and m. mesbahi , `` topology control of dynamic networks in the presence of local and global constraints , '' in _ proc .
( icra ) _ , anchorage , ak , may 2010 , pp . 27182723 .
i. w. mustika , k. yamamoto , h. murata , and s. yoshida , `` potential game approach for spectrum sharing in distributed cognitive radio networks , '' _ ieice trans .
e93-b , no .
32843292 , dec . 2010 .
j. neel , r. m. buehrer , b. h. reed , and r. p. gilles , `` game theoretic analysis of a network of cognitive radios , '' in _ proc .
midwest symp .
circuits syst .
( mwscas ) _ , vol . 3 , tulsa , ok , aug .
2002 , pp .
iii409iii412 .
j. o. neel , r. menon , a. b. mackenzie , j. h. reed , and r. p. gilles , `` interference reducing networks , '' in _ proc .
oriented wireless netw .
( crowncom ) _ , orlando , fl , aug .
2007 , pp . 96104 .
j. o. neel and j. h. reed , `` performance of distributed dynamic frequency selection schemes for interference reducing networks , '' in _ proc .
ieee military commun .
( milcom ) _ ,
washington , dc , oct .
2006 , pp . 17 .
j. o. neel , `` analysis and design of cognitive radio networks and distributed radio resource management algorithms , '' ph.d .
dissertation , virginia polytechnic inst .
state univ .
, blacksburg , va , sep .
2006 .
n. nie , c. comaniciu , and p. agrawal , `` a game theoretic approach to interference management in cognitive networks , '' in _ wireless communications , the i m a volumes in mathematics and its applications_. 1em plus 0.5em minus 0.4emnew york , ny : springer new york , 2007 , vol . 143 , pp . 199219 .
j. ortn , j. r. gllego , and m. canales , `` joint route selection and resource allocation in multihop wireless networks based on a game theoretic approach , '' _ ad hoc netw .
_ , vol . 11 , no . 8 , pp .
22032216 , nov .
m. peltomki , j .-
koljonen , o. tirkkonen , and m. alava , `` algorithms for self - organized resource allocation in wireless networks , '' _ ieee trans . veh .
_ , vol .
61 , no . 1 ,
346359 , jan . 2012 .
s. perlaza , e. belmega , s. lasaulce , and m. debbah , `` on the base station selection and base station sharing in self - configuring networks , '' in _ proc . int .
icst conf .
tools ( valuetools ) _ , pisa , italy , oct . 2009 , pp .
71:171:10 .
v. ramaswamy , v. reddy , s. shakkottai , a. sprintson , and n. gautam , `` multipath wireless network coding : an augmented potential game perspective , '' _ ieee / acm trans .
_ , vol .
22 , no . 1 ,
217229 , feb .
v. reddy , s. shakkottai , a. sprintson , and n. gautam , `` multipath wireless network coding : a population game perspective , '' in _ proc .
( infocom ) _ , san diego , ca , mar .
2010 , pp . 19 .
g. scutari , s. barbarossa , and d. palomar , `` potential games : a framework for vector power control problems with coupled constraints , '' in _ proc
. 31st ieee int .
speech signal process .
( icassp ) _ , vol . 4 ,
toulouse , may 2006 , pp .
241244 .
s. secci , j .- l .
rougier , a. pattavina , f. patrone , and g. maier , `` peering equilibrium multipath routing : a game theory framework for internet peering settlements , '' _ ieee / acm trans .
_ , vol . 19 , no . 2 , pp
. 419432 , apr .
2011 . c. singh , a. kumar , and r. sundaresan , `` uplink power control and base station association in multichannel cellular networks , '' in _ proc .
game theory netw .
( gamenets ) _ , istanbul , may 2009 , pp .
4351 . y. song , c. zhang , and y. fang , `` joint channel and power allocation in wireless mesh networks : a game theoretical perspective , '' _ ieee j. sel .
areas commun .
_ , vol . 26 , no . 7 , pp . 11491159 , sep
. 2008 .
v. srivastava , j. neel , a. b. mackenzie , r. menon , l. a. dasilva , j. e. hicks , j. h. reed , and r. p. gilles , `` using game theory to analyze wireless ad hoc networks , '' _ ieee commun . surveys tuts . _ , vol . 7 , no . 4 , pp .
4656 , jan .
2005 .
r. w. thomas , r. s. komali , a. b. mackenzie , and l. a. dasilva , `` joint power and channel minimization in topology control : a cognitive network approach , '' in _ proc .
( icc ) _ , glasgow , jun .
2007 , pp . 65386543 .
tran , p .-
n . tran , and n. boukhatem , `` strategy game for flow / interface association in multi - homed mobile terminals , '' in _ proc .
( icc ) _ , cape town , south africa , may 2010 , pp .
tseng , f .-
chien , d. zhang , r. y. chang , w .- h .
chung , and c. huang , `` network selection in cognitive heterogeneous networks using stochastic learning , '' _ ieee commun .
_ , vol .
17 , no . 12 , pp .
23042307 , dec .
j. n. tsitsiklis , d. p. bertsekas , and m. athans , `` distributed asynchronous deterministic and stochastic gradient optimization algorithms , '' _ ieee trans .
31 , no . 9 , pp . 803812 , sep .
1986 .
z. uykan and r. jantti , `` joint optimization of transmission - order selection and channel allocation for bidirectional wireless links part i : game theoretic analysis , '' _ ieee trans .
wireless commun .
_ , vol . 13 , no . 7 , pp . 40034013 , jul
. 2014 .
, `` joint optimization of transmission - order selection and channel allocation for bidirectional wireless links part ii : algorithms , '' _ ieee trans .
wireless commun .
_ , vol . 13 , no . 7 , pp .
39914002 , jul .
2014 .
j. wang , y. xu , a. anpalagan , q. wu , and z. gao , `` optimal distributed interference avoidance : potential game and learning , '' _ trans .
_ , vol . 23 , no . 4 , pp .
317326 , jun . 2012 .
q. wang , w. yan , and y. shen , `` @xmath480-person card game approach for solving set @xmath481-cover problem in wireless sensor networks , '' _ ieee trans .
_ , vol .
61 , no . 5 , pp .
15221535 , may 2012 . c. wu , h. mohsenian - rad , j. huang , and a. y. wang , `` demand side management for wind power integration in microgrid using dynamic potential game theory , '' in _ proc .
ieee global telecommun .
workshops ( globecom workshops ) _ , houston , tx , dec .
2011 , pp . 11991204 .
q. wu , y. xu , j. wang , l. shen , j. zheng , and a. anpalagan , `` distributed channel selection in time - varying radio environment : interference mitigation game with uncoupled stochastic learning , '' _ ieee trans .
_ , vol .
62 , no . 9 , pp . 45244538 , nov .
2013 .
y. xu , a. anpalagan , q. wu , l. shen , z. gao , and j. wang , `` decision - theoretic distributed channel selection for opportunistic spectrum access : strategies , challenges and solutions , '' _ ieee commun .
surveys tuts .
_ , vol .
15 , no . 4 , pp . 16891713 , 2013 .
y. xu , j. wang , q. wu , a. anpalagan , and y .- d .
yao , `` opportunistic spectrum access in cognitive radio networks : global optimization using local interaction games , '' _ ieee j. sel
. topics signal process . _ ,
vol . 6 , no . 2 , pp
. 180194 , apr .
2012 .
y. xu , j. wang , y. xu , l. shen , q. wu , and a. anpalagan , `` centralized - distributed spectrum access for small cell networks : a cloud - based game solution , '' 2015 .
[ online ] .
available : http://arxiv.org/abs/1502.06670 y. xu , q. wu , l. shen , j. wang , and a. anpalagan , `` opportunistic spectrum access with spatial reuse : graphical game and uncoupled learning solutions , '' _ ieee trans .
wireless commun .
_ , vol . 12 , no . 10 ,
48144826 , oct .
y. xu , q. wu , j. wang , l. shen , and a. anpalagan , `` opportunistic spectrum access using partially overlapping channels : graphical game and uncoupled learning , '' _ ieee trans .
_ , vol .
61 , no . 9 , pp . 39063918 , sep .
y. xu , y. zhang , q. wu , l. shen , and j. wang , `` distributed spectrum access for cognitive small cell networks : a robust graphical game approach , '' 2015 .
[ online ] .
available : http://arxiv.org/abs/1502.06667 k. yamamoto , h. murata , and s. yoshida , `` management of dominant interference in cognitive radio networks , '' in _ proc .
21st ieee int .
. personal , indoor and mobile radio commun .
( pimrc ) workshops _ , istanbul , turkey , sep .
2010 , pp . 451455 .
d. yang , x. fang , and g. xue , `` channel allocation in non - cooperative multi - radio multi - channel wireless networks , '' in _ proc .
( infocom ) _ , orlando , fl , mar .
2012 , pp . 882890 .
q. yu , j. chen , y. fan , x. shen , and y. sun , `` multi - channel assignment in wireless sensor networks : a game theoretic approach , '' in _ proc .
( infocom ) _ , san diego , ca , mar .
2010 , pp . 19 .
e. zeydan , d. kivanc , and u. tureli , `` cross layer interference mitigation using a convergent two - stage game for ad hoc networks , '' in _ proc .
( ciss ) _ , princeton , nj , mar .
2008 , pp . 671675 .
q. zhang , l. guo , c. sun , x. an , and x. chen , `` joint power control and component carrier assignment scheme in heterogeneous network with carrier aggregation , '' _ iet commun . _ , vol . 8 , no . 10 , pp . 18311836 , jul . 2014 .
z. zhang and c. douligeris , `` convergence of synchronous and asynchronous greedy algorithms in a multiclass telecommunications environment , '' _ ieee trans .
_ , vol .
40 , no . 8 , pp . 12771281 , aug .
j. zheng , y. cai , y. liu , y. xu , b. duan , and x. shen , `` optimal power allocation and user scheduling in multicell networks : base station cooperation using a game - theoretic approach , '' _ ieee trans .
wireless commun .
_ , vol . 13 , no . 12 , pp .
69286942 , dec . 2014 .
j. zheng , y. cai , x. chen , r. li , and h. zhang , `` optimal base station sleeping in green cellular networks : a distributed cooperative framework based on game theory , '' _ to appear in ieee trans .
wireless commun . _ , 2015 .
j. zheng , y. cai , y. xu , and a. anpalagan , `` distributed channel selection for interference mitigation in dynamic environment : a game - theoretic stochastic learning solution , '' _ ieee trans .
_ , vol .
63 , no . 9 , pp . 47574762 , nov .
w. zhong , g. chen , s. jin , and k .- k .
wong , `` relay selection and discrete power control for cognitive relay networks via potential game , '' _ ieee trans .
signal process .
_ , vol .
62 , no . 20 , pp .
54115424 , oct .
2014 .
w. zhong , y. xu , and h. tianfield , `` game - theoretic opportunistic spectrum sharing strategy selection for cognitive mimo multiple access channels , '' _ ieee trans .
signal process .
_ , vol .
59 , no . 6 , pp . 27452759 , jun .
2011 .
koji yamamoto received the b.e .
degree in electrical and electronic engineering from kyoto university in 2002 , and the m.e . and ph.d .
degrees in informatics from kyoto university in 2004 and 2005 , respectively . from 2004 to 2005 , he was a research fellow of the japan society for the promotion of science ( jsps ) .
since 2005 , he has been with the graduate school of informatics , kyoto university , where he is currently an associate professor . from 2008 to 2009 , he was a visiting researcher at wireless@kth , royal institute of technology ( kth ) in sweden .
his research interests include the application of game theory , spectrum sharing , and in - band full - duplex communications .
he received the pimrc 2004 best student paper award in 2004 , the ericsson young scientist award in 2006 , and the young researcher s award from the ieice of japan in 2008 .
he is a member of the ieee . | potential games form a class of non - cooperative games where unilateral improvement dynamics are guaranteed to converge in many practical cases .
the potential game approach has been applied to a wide range of wireless network problems , particularly to a variety of channel assignment problems . in this paper ,
the properties of potential games are introduced , and games in wireless networks that have been proven to be potential games are comprehensively discussed .
potential game , game theory , radio resource management , channel assignment , transmission power control |
the effect of quenched randomness on the equilibrium and dynamic properties of macroscopic systems is a subject of great theoretical and practical interest .
it is well known that quenched bond randomness may produce drastic changes on phase transitions depending on the type of the transition @xcite .
thus , symmetry - breaking first - order transitions are converted to second - order phase transitions by infinitesimal bond randomness for spatial dimensionality @xmath0 @xcite and by bond randomness beyond a threshold strength in @xmath1 @xcite , as indicated by general arguments @xcite and in some cases by rigorous mathematical work @xcite .
in particular , this rounding effect of first - order transitions has now been rigorously established in a unified way in low dimensions ( @xmath2 ) including a large variety of types of randomness in classical and quantum spin systems @xcite .
historically , the effects of disorder on phase transitions have been studied in two extreme cases , i.e. in the limits of weak and strong ( near the percolation point ) disorder .
the first important conjecture , known today as the harris criterion @xcite , relates the value of the specific heat exponent @xmath3 in a continuous transition with the expected effects of uncorrelated weak disorder in ferromagnets . according to the harris criterion , for continuous phase transitions with a negative exponent @xmath3
, the introduction of weak randomness is expected to be an irrelevant field and the disordered system to remain in the same universality class . on the other hand ,
the weakly disordered system is expected to be in a different universality class in the case of a pure system having a positive exponent @xmath3 .
pure systems with a zero specific heat exponent ( @xmath4 ) are marginal cases of harris criterion ( since the criterion does not give any information ) and their study , upon the introduction of disorder , has been of particular interest .
the paradigmatic model of the marginal case is , of course , the general random 2d ising model ( random - site , random - bond , and bond - diluted ) and this model has been extensively investigated and debated @xcite .
several recent studies , both analytical ( renormalization group and conformal field theories ) and numerical ( mainly monte carlo ( mc ) simulations ) devoted to this model , have provided very strong evidence in favor of the so - called logarithmic corrections s scenario . according to this ,
the effect of infinitesimal disorder gives rise to a marginal irrelevance of randomness and besides logarithmic corrections , the critical exponents maintain their 2d ising values .
in particular , the specific heat is expected to slowly diverge with a double - logarithmic dependence @xcite .
here , we should mention that there is not full agreement in the literature and a different scenario predicts a negative specific heat exponent @xmath3 leading to a saturating behavior @xcite , with a corresponding correlation length exponent @xmath5 @xcite . in general , a unitary and rigorous physical description of critical phenomena in disordered systems still lacks and certainly , lacking such a description , the study of further models for which there is a general agreement in the the behavior of the corresponding pure cases is very important .
historically , such a suitable candidate for testing the above predictions , that has been also quite extensively studied , is the general 2d @xmath6-state potts model @xcite .
this model includes the ising model ( @xmath7 ) , cases of a pure system having continuous transitions with a positive exponent @xmath3 ( @xmath8 ) and also the large @xmath6-cases ( @xmath9 ) for which one could observe and try to classify the above mentioned softening of first - order transitions in 2d models .
another similarly interesting candidate , not yet as much studied in the random - bond version , is the 2d blume - capel ( bc ) model @xcite .
we may note here that most of the existing literature on the bc model with randomness concerns randomness applied to the crystal field and/or spin glass exchange interactions @xcite .
as it is well known , the pure version of the bc model undergoes an ising - like continuous phase transition to an ordered ferromagnetic phase as the temperature is lowered for crystal - field couplings less than a tricritical value and a first - order transition for larger values of the crystal - field coupling .
therefore , this model provides also the opportunity to study two different and very interesting topics of the above described effects of disorder in critical phenomena , namely the double - logarithmic scenario for the specific heat in the regime where the 2d bc model is in the same universality class with the 2d random - bond ising model and also the softening of the transition in the first - order regime .
recently the present authors @xcite have considered this model and provided strong numerical evidence clarifying two of the above mentioned effects induced in 2d systems by bond randomness . by implementing a two - stage wang - landau ( wl ) approach @xcite , we presented essentially exact information on the 2d bc model under quenched bond randomness . in this investigation
, we found dramatically different critical behaviors of the second - order phase transitions emerging from the first- and second - order regimes of the pure bc model and since , these second - order transitions were found to have different critical exponents , our study indicated an interesting strong violation of universality @xcite .
namely , different sets of critical exponents on two segments of the same critical line appeared to describe the two regimes : still - second - order and ex - first - order . in this paper ,
we extend our earlier work @xcite , by implementing essentially the same two - stage wl approach ( sec .
[ sec:2 ] ) and try to give a more complete picture by concentrating in the weak ( still - second - order ) regime and simulate the model for several disorder strengths and many values of the crystal - field coupling .
the above statement means that , effectively we will restrict our study to moderate values of the crystal field and moderate values of the disorder , intending to observe the frontier between the weak- and the strong - disorder universality classes from the disappearance of the expected 2d random ising universality class behavior .
thus , in sec . [ sec:3 ]
we will produce phase diagram points for the random - bond model but also for the pure model , reporting for the pure case a comparison with existing estimates in the literature .
more generally , in carrying out this project we have also considered the pure 2d bc model for several values of the crystal - field , in the second - order regime , observing its finite - size scaling ( fss ) behavior .
[ sec:4 ] presents such a comparative study between random and pure models concerning the behavior of all thermodynamic parameters used in the traditional fss analysis of mc data .
this study enables us to observe some peculiarities of the pure model , due to the onset of tricriticality , and compare them with the corresponding behavior of the random model .
furthermore , we try to focus , understand and shed light to the extent of universality of the random - bond 2d bc model with the corresponding random - bond ising model , for which the scenario of logarithmic corrections seems to be the strongest option in the current literature @xcite .
the prediction of the range of such universality is far from trivial and the two regimes ( weak and strong ) have many dissimilarities which are also reflected in the ground - state structure , as further discussed below .
the attempt to estimate the range of the above mentioned universality is accomplished by the novel idea which assumes the truth of the double logarithmic scenario for the specific heat in a suitable restricted range .
this is presented in sec .
[ sec:5 ] together with some further crucial observations concerning the behavior of the strong disorder regime , i.e. the regime where the ising class universality does not apply and the system has a rather complex and yet not understood behavior .
our conclusions are summarized in sec .
[ sec:6 ] .
the ( pure ) bc model @xcite is defined by the hamiltonian @xmath10 where the spin variables @xmath11 take on the values @xmath12 , or @xmath13 , @xmath14 indicates summation over all nearest - neighbor pairs of sites , and @xmath15 is the ferromagnetic exchange interaction .
the parameter @xmath16 is known as the crystal - field coupling and to fix the temperature scale we set @xmath17 and @xmath18 . as it is well known , this model has been analyzed , besides the original mean - field theory @xcite , by a variety of approximations and numerical approaches .
these include the real space renormalization group , mc simulations , and mc renormalization - group calculations @xcite , @xmath19-expansion renormalization groups @xcite , high- and low - temperature series calculations @xcite , a phenomenological fss analysis using a strip geometry @xcite , and , finally , a recent two - parameter wl sampling in rather small lattices of linear sizes @xmath20 @xcite . as mentioned already in the introduction the phase diagram of the model consists of a segment of continuous ising - like transitions at high temperatures and low values of the crystal field which ends at a tricritical point , where it is joined with a second segment of first - order transitions between ( @xmath21 ) and ( @xmath22 ) . the model given by eq .
( [ eq:1 ] ) is studied here on the square lattice and will be referred to as the pure bc model . however , our main focus , on the other hand , is the case with bond disorder given by the bimodal distribution @xmath23\;;\\ \nonumber & \frac{j_{1}+j_{2}}{2}=1\;;\;\;j_{1}>j_{2}>0\;;\;\;r=\frac{j_{2}}{j_{1}}\;,\end{aligned}\ ] ] so that @xmath24 reflects the strength of the bond randomness .
the resulting quenched disordered ( random - bond ) version of the hamiltonian defined in eq .
( [ eq:1 ] ) reads now as @xmath25 a first comparative study between the two versions ( pure and random ) of the 2d bc model has been already presented in our earlier paper @xcite for two values of the crystal - field coupling corresponding to the second - order ( @xmath26 ) and first - order ( @xmath27 ) regimes of the pure model where for the random version the disorder strength @xmath28 was chosen in both cases . in the next sections our study on the 2d bc model is extended to several values of the crystal field and disorder ( listed in the table of sec .
[ sec:3 ] ) .
details of our simulations are summarized in the next section together with an up to date brief sketch of our entropic scheme . in the last few years
we have used an entropic sampling implementation of the wl algorithm @xcite to study some simple @xcite , but also some more complex systems @xcite
. one basic ingredient of this implementation is a suitable restriction of the energy subspace for the implementation of the wl algorithm .
this was originally termed as the critical minimum energy subspace ( crmes ) restriction @xcite and it can be carried out in many alternative ways , the simplest being that of observing the finite - size behavior of the tails of the energy probability density function ( e - pdf ) of the system @xcite .
complications that may arise in random systems can be easily accounted for by various simple modifications that take into account possible oscillations in the e - pdf and expected sample - to - sample fluctuations of individual disorder realizations . in our recent papers @xcite , we have presented details of various sophisticated routes for the identification of the appropriate energy subspace @xmath29 for the entropic sampling of each random realization . in estimating the appropriate subspace from a chosen pseudocritical temperature one should be careful to account for the shift behavior of other important pseudocritical temperatures and extend the subspace appropriately from both low- and high - energy sides in order to achieve an accurate estimation of all finite - size anomalies .
of course , taking the union of the corresponding subspaces , insures accuracy for the temperature region of all studied pseudocritical temperatures .
the up to date version of our implementation uses a combination of several stages of the wl process .
first , we carry out a starting ( or preliminary ) multi - range ( multi - r ) stage , in a very wide energy subspace .
this preliminary stage is performed up to a certain level of the wl random walk .
the wl refinement is @xmath30 , where @xmath31 is the density of states ( dos ) and we follow the usual modification factor adjustment @xmath32 and @xmath33 .
the preliminary stage may consist of the levels : @xmath34 and to improve accuracy the process may be repeated several times .
however , in repeating the preliminary process and in order to be efficient , we use only the levels @xmath35 after the first attempt , using as starting dos the one obtained in the first random walk at the level @xmath36 . from our experience , this practice is almost equivalent of simulating the same number of independent wl random walks . also in our recent studies we have found out that is much more efficient and accurate to loosen up the originally applied very strict flatness criteria @xcite .
thus , a variable flatness process starting at the first levels with a very loose flatness criteria and assuming at the level @xmath37 the original strict flatness criteria is now days used . after the above described preliminary multi - r stage , in the wide energy subspace
, one can proceed in a safe identification of the appropriate energy subspace using one or more alternatives outlined in refs .
@xcite . in random systems , where one needs to simulate many disorder realizations ,
it is also possible and advisable to avoid the identification of the appropriate energy subspace separately for each disorder realization by extrapolating from smaller lattices and/or by prediction from preliminary runs on small numbers of disorder realizations . in any case
, the appropriate subspaces should be defined with sufficient tolerances . in our implementation
we use such advance information to proceed in the next stages of the entropic sampling .
the process continues in two further stages ( two - stage process ) , using now mainly high iteration levels , where the modification factor is very close to unity and there is not any significant violation of the detailed balance condition during the wl process .
these two stages are suitable for the accumulation of histogram data ( for instance energy - magnetization histograms ) , which can be used for an accurate entropic calculation of non - thermal thermodynamic parameters , such us the order parameter and its susceptibility @xcite . in the first ( high - level )
stage , we follow again a repeated several times ( typically @xmath38 ) multi - r wl approach , carried out now only in the restricted energy subspace .
the wl levels may be now chosen as @xmath39 and as an appropriate starting dos for the corresponding starting level the average dos of the preliminary stage at the starting level may be used .
finally , the second ( high - level ) stage is applied in the refinement wl levels @xmath40 ( typically @xmath41 ) , where we usually test both an one - range ( one - r ) or a multi - r approach with large energy intervals . in the case of the one - r approach
we have found very convenient and in most cases more accurate to follow the belardinelli and pereyra @xcite adjustment of the wl modification factor according to the rule @xmath42 . finally , it should be also noted that by applying in our scheme a separate accumulation of histogram data in the starting multi - r stage ( in the wide energy subspace ) offers the opportunity to inspect the behavior of all basic thermodynamic functions in an also wide temperature range and not only in the neighborhood of the finite - size anomalies .
the approximation outside the dominant energy subspace is not of the same accuracy with that of the restricted dominant energy subspace but is good enough for the observation of the general behavior and provides also a route of inspecting the degree of approximation .
the above described numerical approach was used to estimate the properties of a large number of @xmath43 bond disorder realizations , for lattice sizes @xmath44 for all crystal fields and disorder strengths used in this paper , with the exception of the case @xmath45 and @xmath46 , where @xmath47 disorder realizations were simulated for lattice sizes @xmath48 . for reference and contrast ,
the pure - system s properties were also obtained by the same implementation , simulating in each case at least @xmath49 independent runs .
we close this outline of our numerical scheme with some comments concerning statistical errors and disorder averaging . even for the larger lattice size studied here ( @xmath50 ) , and depending on the thermodynamic parameter
, the statistical errors of the wl method were found to be of reasonable magnitude and in some cases to be of the order of the symbol sizes , or even smaller .
this was true for both the pure version and the individual random - bond realizations .
however , since the nature of the present study is qualitative , not aiming to an accurate exponent estimation , the wl errors will not be presented in our figure illustrations .
we also note that , for the random - bond version mainly the averages over the disorder realizations , denoted as @xmath51_{av}$ ] , will be considered in the text and their finite - size anomalies , denoted as @xmath51_{av}^{\ast}$ ] , will be used in our fss attempts . due to very large sample - to - sample fluctuations ,
mean values of individual maxima ( @xmath52_{av}$ ] ) have not been used in this study except for illustrative purposes , as in the case @xmath45 , @xmath46 presented in our last section .
( a ) simultaneous fitting of the six pseudocritical temperature defined in the text for the random - bond version for @xmath45 and @xmath53 .
the fitting range is @xmath44 .
( b ) the same as above simultaneous fitting of the six pseudocritical temperatures now for the pure model at @xmath54 .
the solid line shows our estimate of the critical temperature @xmath55 ( with the dotted lines indicating its errors barriers ) and the dashed line the estimate @xmath56 given in ref .
@xcite.,width=8 ] this section presents , as mentioned in the introduction , phase diagram points for the pure and random - bond models and in the case of the pure model compare the corresponding phase diagram points with the existing estimates in the literature .
this gives also the opportunity to observe the reliability of our numerical approach .
following the practice of our earlier paper @xcite , we estimate phase diagram points by fitting our data to the expected power - law shift behavior @xmath57 of several pseudocritical temperatures . the traditionally used specific heat and magnetic susceptibility peaks , as well as
, the peaks corresponding to the following logarithmic derivatives of the powers @xmath58 of the order parameter with respect to the inverse temperature @xmath59 @xcite , @xmath60 and the peak corresponding to the absolute order - parameter derivative @xmath61 will be implemented for a simultaneous fitting attempt .
phase diagrams of the pure 2d bc model and its random version at the disorder strength @xmath62 .
the inset illustrates the crossing of the phase boundaries including an approximate estimate of the crossing point.,width=8 ] such simultaneous fitting attempts are presented in fig .
[ fig:1 ] .
in particular fig .
[ fig:1](a ) presents the shift behavior for the random - bond 2d bc model in the weak disorder case @xmath53 at the value @xmath45 and fig .
[ fig:1](b ) illustrates the simultaneous fitting attempted for the pure model at @xmath54 , further discussed at the end of sec .
[ sec:4b ] .
as noted already , the data fitted for the random version of the model are only those of the pseudocritical temperatures of the peaks of the averaged over disorder thermodynamic parameters , as indicated in the figure , where the asterisk denotes the peak of the averaged over disorder parameter @xmath51^{\ast}_{av}$ ] .
the alternative route of using averages of individual sample parameters gives almost identical estimates .
however , in cases of strong lack of self - averaging , very large sample - to - sample fluctuations may be present . for simplicity , in all fitting attempts , the whole range @xmath44 has been used . following this practice for the pure and random version of the 2d bc several phase diagram points
were produced .
figure [ fig:2 ] shows the resulting phase diagrams using our phase diagram points for the pure 2d bc model and its random - bond version for the disorder strength @xmath28 .
the tricritical point shown is taken from the estimate given by beale @xmath63=@xmath64 @xcite .
for the disorder strength @xmath62 the points of the phase diagram were chosen with the intension to be able to approximately locate the emergence of the enhancement of ferromagnetic order observed in our earlier study @xcite in the ex - first - order regime at @xmath27 , where we found a considerable increase of the critical temperature by @xmath65 .
a microscopic explanation of this phenomenon , based on the preference @xmath66 states to the strong - coupling connectivity sites has been given in ref .
@xcite ( see also ref .
@xcite ) , as a microsegregation process , due to quenched bond randomness , that evolves continuously within the ferromagnetic and paramagnetic phases . the inset of fig .
[ fig:2 ] shows that phase diagram of the random version crosses the diagram of the pure model at approximately the crossing point @xmath67 , well before the tricritical point .
the approximation of this point was obtained by interpolation using the four points of the phase diagram in the range @xmath68 .
table [ tab:1 ] summarizes the phase diagram points obtained in this paper , by the above described traditional ffs method for both the pure and random version of the 2d bc model for disorder strengths @xmath53 , @xmath62 , and @xmath69 , together with corresponding phase diagram points given by beale @xcite for the pure model . note that , in silva _
et al . _
@xcite one can find an analogous table for the pure 2d bc model including the phase diagram points given by beale @xcite and the points produced in their two - parametric wl sampling @xcite .
it can be seen that there is an excellent coincidence of our points with those of beale @xcite .
the points of beale @xcite are based on the very accurate phenomenological fss scheme using a strip geometry @xcite , whereas our points are obtained via the present simultaneous fss analysis , based on lattices with linear sizes @xmath44 .
the points of silva _ et al . _
@xcite are based in much smaller lattices of the two - parameter wl sampling ( linear sizes @xmath20 ) .
however , in the case @xmath54 our phase diagram point does not agree , within errors , with that of beale and for this reason our estimation with a rather generous error bars ( shown also on the panel ) has been illustrated in fig .
[ fig:1](b ) .
the rest of table [ tab:1 ] contains our estimates for the random - bond versions .
one may note from this table that for small values of @xmath16 for instance @xmath45 the corresponding critical temperatures decrease as the disorder becomes stronger ( compare the three cases of disorder : @xmath53 , @xmath62 , and @xmath69 at this value of the crystal field ) . on the other hand , for @xmath70 the trend
is reversed as can be seen by comparing the corresponding three cases of disorder .
apparently this is a kind of reflection of the phenomenon of the enhancement of ferromagnetic order which appears to influence the geometry of the critical surface for the 2d random - bond bc model . [ cols="<,^,^,^,^,^,^",options="header " , ]
as pointed out in the introduction , in our recent investigation of the 2d random - bond bc model @xcite we found dramatically different critical behaviors of the second - order phase transitions emerging from the first- and second - order regimes of the pure model .
namely , different sets of critical exponents on two segments of the same critical line appeared to describe the two regimes : the still - second - order and ex - first - order regimes .
the study in ref .
@xcite was carried out for two values of the crystal - field coupling corresponding to the second - order ( @xmath26 ) and first - order ( @xmath27 ) regimes of the pure model and for the random version the disorder strength @xmath62 was chosen in both cases .
the strong violation of universality observed appeared to be the result of the softening of the first - order transition due to bond - randomness .
specifically , it was concluded that the new strong - disorder universality class is well described by a correlation length exponent in the range @xmath71 , and exponent ratios @xmath72 and @xmath73 very close to the ising values @xmath74 and @xmath75 , respectively @xcite .
the above weak universality @xcite seems to be valid between the ising - like continuous transitions of the random - bond 2d bc model for small values of @xmath16 ( @xmath76 in ref .
@xcite ) and continuous transitions belonging to the strong - disorder universality class .
therefore , the strong - disorder universality class may be characterized by the above distinct value of the correlation length exponent and a strong saturation of the specific heat .
qualitatively this saturating behavior is quite instructive and for illustrative reasons is reproduced here in fig .
[ fig:3 ] .
this figure contrasts , at @xmath27 , the specific heat s finite - size behavior of the pure 2d bc model ( first - order regime ) and two disordered cases corresponding to disorder strengths @xmath77 and @xmath62 .
the saturation of the specific heat is very clear in both cases of the disorder strength .
it is of interest to point out here that these findings for the strong - disorder universality class appear to be fully compatible with the classification of phase transitions in disordered systems proposed recently by wu @xcite . according to this classification
the strong - disorder transition is expected to be inhomogeneous and percolative with an expected exponent of the order @xmath78 @xcite .
furthermore , it has been suggested to us by wu @xcite , that the strong lack of self - averaging of this transition stems from the above properties .
this violation of self - averaging , together with the strong finite - size effects , make the systematic mc approach of the strong - disorder regime very demanding , if not impractical . on the other hand ,
the weak regime ( or ising universality regime ) suffers a much weaker lack of self - averaging , by at least a factor of @xmath79 @xcite , and a smooth behavior is observed at moderate lattice sizes .
thus , aiming here to observe , even approximately , the extent of the involved universality classes we carried out our study at moderate values of the crystal field and disorder , and found a behavior quite convincing from which the frontier of the strong universality class can be estimated by observing the disappearance of the expected 2d random ising universality class .
let us now proceed with the analysis of our numerical data for the disorder strengths and crystal fields given in table [ tab:1 ] and observe and contrast their fss behavior with that of the pure model .
starting this comparative study with the fss of the specific heat maxima ( using for the random - bond version at the strength @xmath62 the corresponding quantity averaged over disorder , i.e. @xmath80_{av}^{\ast}$ ] ) , we present in fig . [ fig:4 ] fitting attempts for the same range of @xmath16 for the pure model [ fig .
[ fig:4](a ) ] and the random - bond version [ fig .
[ fig:4](b ) ] .
behavior of the random - bond 2d bc model at @xmath27 from ref .
illustration of the divergence of the specific heat of the pure model ( first - order regime ) and the clear saturation of the specific heat for the random - bond ( open symbols ) 2d bc model for two disorder strengths in a log - log scale.,width=8 ] as indicated by the scales in the x - axis and the functions in the corresponding panels , the expected ising logarithmic divergence has been assumed for the pure model @xmath81 , whereas the double - logarithmic divergence @xmath80_{av}^{\ast}=c_{1}+c_{2}\ln(\ln l)$ ] has been assumed for the random version .
although , it is very difficult to irrefutably distinguish between a double - logarithmic divergence and a very weak power - law divergence , the theoretically well - grounded double - logarithmic scenario applies very well @xcite and this fact can be observed also now for more values of @xmath16 in fig .
[ fig:4](b ) .
there are some further features one can observe from fig .
[ fig:4 ] .
first , for the pure model , and in the range @xmath82 , we observe a sudden change in the behavior of the specific heat peaks as we approach the tricritical point which is apparently a strong cross - over effect . for the random version and the same values of @xmath16
no such strong effects are noticeable and most probably the general softening effects of bond randomness extends also to the expected cross - over phenomena between the two different universality classes of the random 2d bc model . from fig .
[ fig:4](b ) the slopes of the double - logarithmic fittings appear to obey a rather sensible decreasing tendency from which we try in the next section to locate the frontier of strong - disorder universality class .
fss of the specific heat maxima for the pure and random - bond 2d bc model at the same values of @xmath16 .
( a ) pure model : illustration of linear fittings assuming the expected ising logarithmic divergence .
( b ) random model : illustration of linear fittings assuming a double logarithmic divergence . note the steady fall of the double logarithmic amplitude @xmath83.,width=8 ] a second interesting comparison follows now in fig .
[ fig:5 ] , where again fig .
[ fig:5](a ) presents the fss of the susceptibility maxima for the pure model , whereas fig .
[ fig:5](b ) corresponds to the random - bond version at @xmath62 .
the influence of the exceptionally large fluctuations in the order parameter of the pure model , as we approach the tricritical point , is now reflected in the effective exponent ratio @xmath72 , estimated by the simple power law @xmath84 .
these large fluctuations are a known peculiarity of the pure model near tricriticality @xcite .
the effective value of the exponent is now closer to the expected value at the tricritical point @xmath85
@xcite than to the value of the ising universality class @xmath86 . for this reason a separate power - law fitting has been applied for the value @xmath54 in fig .
[ fig:5](a ) .
note here that in fig .
[ fig:5](a ) the simultaneous fitting is applied to the first six values of @xmath87 , whereas in fig .
[ fig:5](b ) all values of @xmath88 are used .
however , even so , the comparison of the two panels of fig .
[ fig:5 ] , points out the much better limiting behavior of the random version towards the expected ising value @xmath86 .
therefore , we may convincingly suggest that , for moderate values of crystal field and disorder , the ising universality scenario ( with possible logarithmic corrections ) is well obtained in the disordered case
. figures [ fig:6 ] and [ fig:7 ] illustrate further alternative routes , used commonly in traditional fss analysis , that provide clear evidence to the above suggestion namely that the weak - disorder version belong to the 2d ising class for suitable moderate values of disorder and crystal - field coupling .
noteworthy is the fact that the estimation of the critical exponents via the traditional fss , such as that shown in figs .
[ fig:5 ] - [ fig:7 ] for the random version , yields estimates very close to the expected values , i.e. the exponents of the 2d ising model . on the other hand , as pointed also earlier , for the pure model , as we approach the tricritical point ( @xmath54 ) the effective exponents remind more those expected at the tricritical point than those of the 2d ising model .
fss behavior of the susceptibility maxima for the pure and random - bond 2d bc model at the same values of @xmath16 .
( a ) pure model : simultaneous fitting to a simple power law for the first five values of @xmath16 and a separate fitting close to the tricritical point for @xmath54 .
( b ) random model : simultaneous fitting of the averaged susceptibility peaks for all values of @xmath16 . note the better behavior for the random model and the improved estimation of the exponent ratio @xmath72.,width=8 ] this is particularly true for the correlation length s exponent estimated in fig .
[ fig:1](b ) by the simultaneous fitting of the six pseudocritical temperatures .
as can be seem from this figure the estimate for the exponent @xmath89 , in the case @xmath54 , is closer to the expected tricritical value @xmath90
@xcite , than to the 2d ising value @xmath91 .
( a ) fss analysis of the three logarithmic derivatives ( @xmath58 ) of the order parameter with respect to temperature for a particular crystal field @xmath92 and disorder strength @xmath62 .
( b ) traditional fss analysis of the order parameter at the estimated critical temperature for the same value of crystal field and disorder strength , as in panel ( a).,width=8 ]
from fig . [ fig:4](a ) one can observe the expected ising logarithmic divergence of the specific heat maxima .
avoiding the value @xmath54 , which suffers from strong cross - over effects , we attempted to estimate the tricritical value of the crystal field by fitting the decreasing logarithmic amplitudes @xmath93 to a suitable power law , as shown in fig .
[ fig:8 ] .
this may appear a naive or questionable idea , since the behavior of specific heat data is the achilles heel of fss analysis .
yet , fig .
[ fig:8 ] shows that besides the large fluctuations ( errors ) in logarithmic amplitudes @xmath93 , one could approximately estimate the tricritical crystal field @xmath94 , as shown in the panel of fig .
[ fig:8 ] . from fig .
[ fig:4](b ) we observe again that , the specific heat maxima , corresponding now to the random - bond model at the disorder strength @xmath62 , are better - matched to the double - logarithmic divergence than the corresponding specific heat maxima of the pure model to a simple logarithmic divergence
. therefore , it appears realistic to try to obtain the multicritical point ( more precisely the value of the crystal field @xmath95 where we expect the emergence of strong - disorder regime at a given disorder strength ) , where the 2d random - bond ( bc ) ising universality class meets the strong - disorder universality class , by fitting the decreasing double - logarithmic amplitudes to a similar power law .
simultaneous fitting to a simple power law of the averaged peaks corresponding to the absolute order - parameter derivative for three values of @xmath16 and disorder strength @xmath96 .
the fitting provides an estimate of the exponent @xmath97 and as indicated in the panel an alternative estimate for the exponent ratio @xmath73 , by assuming @xmath91.,width=8 ] figure [ fig:9 ] presents now these fittings and also the estimated multicritical values of the crystal field @xmath95 for the three disorder strengths considered in this paper .
the statistical errors of the corresponding double logarithmic amplitudes @xmath93 are seen to be quite smaller , compared to the corresponding amplitudes of the pure model , and the estimated values of the values of the multicritical field , shown in the panel of fig .
[ fig:9 ] , appear more convincing . from their general trend
we observe that as we increase the disorder strength the frontier of the strong - disorder universality class moves to lower values of the crystal field , namely : ( @xmath53 , @xmath98 ) , ( @xmath62 , @xmath99 ) , and ( @xmath69 , @xmath100 ) .
this behavior seems sensible and in our opinion reflects the competition between the ferromagnetic random interactions with the crystal field , giving a kind of destabilization of the usual ising - like ferromagnetic order .
a similar ground - state reflection of this competition is a ground - state structure of unsaturated ferromagnetic ground states discussed in our recent paper @xcite .
conversely , as disorder strength is decreased , @xmath95 approaches @xmath101 , as expected .
as pointed out earlier , the strong lack of self - averaging , together with possible strong finite - size effects , make the mc approach to the strong - disorder regime a very difficult task .
the self - averaging properties along the two segments ( ex - first - order and still - second - order ) of the critical line were observed and discussed in ref .
it was shown in this paper that the usual finite - size measure @xcite of relative variance @xmath102_{av}^{2}$ ] , where @xmath103_{av}-[x]_{av}^{2}$ ] ( and @xmath104 is the susceptibility maxima ) , exhibits lack of self - averaging in both cases of marginal second - order transition and the transition in the strong - disorder or ex - first - order regime .
in particular , it was shown @xcite that the case studied ( @xmath27 ) of the ex - first - order segment gives a much larger effect when compared with the still - second - order regime at @xmath26 and the same disorder strength @xmath62 by a factor of @xmath105 . an approximate estimation of the tricritical value of the crystal field by fitting the decreasing logarithmic amplitudes @xmath83 of the pure model at suitable values of the crystal - field coupling to a suitable power law shown in the figure.,width=8 ] therefore , phase diagram points very close to the frontier of the strong - disorder regime may be the worst cases to study , because , besides the very strong lack of self - averaging one may also expect large finite - size and cross - over effects .
however , such cases are elucidatory not only for observing the intrinsic difficulties but also for giving us the opportunity to reflect on possible links with basic properties of the system as for instance the ground - state structure . thus , fig .
[ fig:10 ] illustrates two important characteristics of the case @xmath45 and @xmath46 .
in particular , fig .
[ fig:10](a ) shows the huge sample - to - sample fluctuations in all the pseudocritical temperatures used in this paper .
the simulation here was extended to larger lattices ( @xmath48 ) and the number of realizations studied was @xmath47 , to be compared with @xmath43 realizations studied in previous sections ( weak - regime ) .
besides the enormous fluctuations , the number of @xmath47 realizations looks quite insufficient and strong finite - size effects , reflected as expected in the shift behavior of the system , produce a completely unsettled behavior .
figure [ fig:10](b ) illustrates an unusual deviating behavior between the finite - size behavior of the maxima of the averaged specific heat curves ( @xmath80_{av}^{\ast}$ ] ) and the finite - size behavior of the average of individual maxima ( @xmath106_{av}$ ] ) , which as shown suffer large sample - to - sample fluctuations .
the behavior here resembles in many aspects the well known and still challenging specific heat behavior of the 3d random - field ising model @xcite .
however , a very clear tendency of @xmath80_{av}^{\ast}$ ] for a saturating behavior is observed in fig .
[ fig:10](b ) and we may speculate that this saturating behavior is the correct asymptotic behavior for both maxima shown in fig .
[ fig:10](b ) , although the behavior of @xmath106_{av}$ ] will settle down only in very large lattice sizes , when the influence of the strong finite - size effects on the individual maxima will diminish .
estimation of multicritical values of the crystal field , where the 2d random - bond ( bc ) ising universality class meets the strong - disorder universality class .
the decreasing double logarithmic amplitudes @xmath83 of the random version has been fitted to a power law shown and the estimates for the three disorder strengths are shown in the panel.,width=8 ] the above illustrations open the possibility that the case @xmath45 and @xmath46 is very close to the frontier of the strong - universality class .
this is in accordance with the trend observed in the previous section and can be further supported by reproducing here some aspects of the ground - state structure of the 2d random - bond bc model . from fig .
[ fig:11 ] , reproduced here from ref .
@xcite , one can see that approximately at this point ( @xmath107 ) , the system departs from the ferromagnetic ground state and an unsaturated ground state is produced , which is further enhanced with vacant sites ( @xmath108 ) as we increase the disorder strength . in the presence of bond randomness the competition between the ferromagnetic interactions with the crystal field results in a destabilization of the ferromagnetic ground state .
depending on the realization , weak clusters exist in @xmath109 and their points are frozen in the @xmath108 state .
this is an interesting subject , which is presently under further consideration in both 2d and 3d by the present authors .
the novel behavior illustrated in fig . [ fig:10 ] appears now as a consequence of the onset of the unsaturated ground state at @xmath110 which is thus related with the critical behavior of the 2d random - bond bc model .
the presented calculation of the ground states has be carried out in polynomially bounded computing time by mapping the system into a network and searching for a minimum cut by using a maximum flow algorithm ( see for instance ref .
@xcite ) and can be easily extended to large lattices and also to the 3d bc random - bond model .
finite - size behavior for the case @xmath45 and @xmath46 .
( a ) behavior of the averaged pseudocritical temperatures , corresponding to individual maxima , with an illustration of the huge sample - to - sample fluctuations .
( b ) behavior of the maxima of the averaged specific heat curves ( @xmath80_{av}^{\ast}$ ] ) and the average of individual maxima ( @xmath106_{av}$ ] ) with their large sample - to - sample fluctuations.,width=8 ]
by carrying out an extensive two - stage wang - landau entropic sampling of both the pure and the random - bond 2d blume - capel model we have produced phase diagram points for several disorder strengths . also for the pure model we found an excellent coincidence of our points with those of beale @xcite . for a particular disorder strength ( @xmath62 ) we found that , as a result of the enhancement of ferromagnetic order , the phase diagram of the random version crosses that of the pure model at approximately the point @xmath67 , well before the tricritical point .
the critical properties of the pure model were compared and contrasted to those of the random model and for moderate values of the crystal field and disorder we found that the ising universality scenario ( with possible logarithmic corrections ) is well obtained in the case of the random version .
furthermore , accepting in this range of couplings the assumption of the double logarithmic scenario for the specific heat we estimated multicritical points , where the 2d random - bond ( bc ) ising universality class meets the strong - disorder universality class .
the behavior of the strong disorder regime was also critically discussed . the case @xmath45 and @xmath46
was extensively studied and based on the observed behavior and on the ground - state observations we suggested that this case , most likely , lies on the frontier between the weak- and strong - disorder universality classes , suffering exceptionally strong finite - size effects and a strong lack of self - averaging . in conclusion ,
the present paper investigated some difficult and important aspects of the random - bond 2d blume - capel model .
it has been pointed out that the behavior of this system is very interesting and in particular the strong - disorder regime may include many further challenges and open problems that , at the moment , are not fully understood . in our opinion
, this is a rather complex subject deserving further research .
harris , j. phys .
c * 7 * , 1671 ( 1974 ) .
berker , phys .
b * 42 * , 8640 ( 1990 ) .
m. aizenman and j. wehr , phys .
lett . * 62 * , 2503 ( 1989 ) ; * 64 * , 1311(e ) ( 1990 ) .
k. hui and a.n .
berker , phys .
lett . * 62 * , 2507 ( 1989 ) ; * 63 * , 2433(e ) ( 1989 ) .
berker , physica a * 194 * , 72 ( 1993 ) .
a. falicov and a.n .
berker , phys .
lett . * 76 * , 4380 ( 1996 ) .
greenblatt , m. aizenman , and j.l .
lebowitz , phys .
103 * , 197201 ( 2009 ) ; physica a ( to appear ) , also at arxiv:0912.1251 .
fytas , a. malakis , and i.a .
hadjiagapiou , j. stat .
( 2008 ) p11009 .
g. grinstein and a. luther , phys .
b * 13 * , 1329 ( 1976 ) .
r. fisch , j. stat .
phys . * 18 * , 111 ( 1978 ) .
v.s . dotsenko and v.s .
dotsenko , sov .
jetp lett .
* 33 * , 37 ( 1981 ) .
shalaev , sov .
solid state * 26 * , 1811 ( 1984 ) .
r. shankar , phys .
lett . * 58 * , 2466 ( 1987 ) .
ludwig , nucl .
b * 285 * , 97 ( 1987 ) .
wang , w. selke , v.s .
dotsenko , and v.b .
andreichenko , physica a * 164 * , 221 ( 1990 ) .
kim and a. patrascioiu , phys .
lett . * 72 * , 2785 ( 1994 ) . v. dotsenko , m. picco , and p. pujol , nucl . phys .
b * 455 * , 701 ( 1995 ) .
aaro reis , s.l.a .
de queiroz , and r.r . dos santos , phys . rev .
b * 54 * , r9616 ( 1996 ) .
ballesteros , l.a .
fernndez , v. martn - mayor , a. muoz sudupe , g. parisi , and j.j .
ruiz - lorenzo , j. phys .
a * 30 * , 8379 ( 1997 ) .
w. selke , l.n .
shchur , and o.a .
vasilyev , physica a * 259 * , 388 ( 1998 ) . g. mazzeo and r. khn , phys .
e * 60 * , 3823 ( 1999 ) .
martins and j.a .
plascak , phys . rev .
e * 76 * , 012102 ( 2007 ) .
m. hasenbusch , f.p .
toldin , a. pelissetto , and e. vicari , phys .
e * 78 * , 011110 ( 2008 ) .
a. gordillo - guerrero , r. kenna , and j.j .
ruiz lorenzo , aip conf .
proc . * 1198 * , 42 ( 2009 ) .
chayes , l. chayes , d.s .
fisher , and t. spencer , phys .
lett . * 57 * , 2999 ( 1986 ) .
ludwig and j.l .
cardy , nucl .
b * 285 * , 687 ( 1987 ) .
s. chen , a.m. ferrenberg , and d.p .
landau , phys .
e * 52 * , 1377 ( 1995 ) .
s. wiseman and e. domany , phys .
e * 51 * , 3074 ( 1995 ) .
j.l . jacobsen and j. cardy ,
b * 515 * 701 ( 1998 ) .
m. picco , phys .
b * 54 * , 14930 ( 1996 ) ; phys . rev
. lett . * 79 * , 2998 ( 1997 ) .
m. blume , phys .
141 * , 517 ( 1966 ) .
capel , physica ( utr . ) * 32 * , 966 ( 1966 ) ; * 33 * , 295 ( 1967 ) ; * 37 * , 423 ( 1967 ) .
m. kaufman and m. kanner , phys .
b * 42 * , 2378 ( 1990 ) .
i. puha and h.t .
diep , j. magn .
magn . mater .
* 224 * , 85 ( 2001 ) .
o.d . rodriguez salmon and j.r .
tapia , arxiv : 0910.3202 .
zelik and a.n .
berker , phys .
e * 78 * , 031104 ( 2008 ) .
a. malakis , a.n .
berker , i.a .
hadjiagapiou , and n.g .
fytas , phys . rev .
e * 79 * , 011125 ( 2009 ) .
f. wang and d.p .
landau , phys .
rev . lett . * 86 * , 2050 ( 2001 ) .
f. wang and d.p .
landau , phys .
e * 64 * , 056101 ( 2001 ) .
a. malakis , a. peratzakis , and n.g .
fytas , phys .
e * 70 * , 066128 ( 2004 ) .
a. malakis , s.s .
martinos , i.a .
hadjiagapiou , n.g .
fytas , and p.
kalozoumis , phys .
e * 72 * , 066120 ( 2005 ) .
a. malakis and n.g .
fytas , phys .
e * 73 * , 016109 ( 2006 ) .
fytas , a. malakis , and k. eftaxias , j. stat .
( 2008 ) p03015 .
landau , phys .
* 28 * , 449 ( 1972 ) ; a.n .
berker and m. wortis , phys . rev .
b * 14 * , 4946 ( 1976 ) ; m. kaufman , r.b .
griffiths , j.m .
yeomans , and m. fisher , _
ibid_. * 23 * , 3448 ( 1981 ) ; w. selke and j. yeomans , j. phys . a * 16 * , 2789 ( 1983 ) ; d.p .
landau and r.h .
swendsen , phys .
b * 33 * , 7700 ( 1986 ) ; j.c .
xavier , f.c .
alcaraz , d. pena lara , and j.a .
plascak , _ ibid . _ * 57 * , 11575 ( 1998 ) .
stephen and j.l .
mccoley , phys .
lett . * 44 * , 89 ( 1973 ) ; t.s .
chang , g.f .
tuthill , and h.e .
stanley , phys .
b * 9 * , 4482 ( 1974 ) ; g.f . tuthill , j.f .
nicoll , and h.e .
stanley , _ ibid .
_ * 11 * , 4579 ( 1975 ) ; f.j .
wegner , phys . lett . *
54a * , 1 ( 1975 ) .
p.f . fox and a.j .
guttmann , j. phys .
c * 6 * , 913 ( 1973 ) ; t.w .
burkhardt and r.h .
swendsen , phys .
b * 13*. 3071 ( 191976 ) ; w.j . camp and j.p .
van dyke , _ ibid .
_ * 11 * , 2579 ( 1975 ) .
p. nightingale , j. appl . phys . *
53 * , 7927 ( 1982 ) .
beale , phys .
b * 33 * , 1717 ( 1986 )
silva , a.a .
caparica , and j.a .
plascak , phys .
e * 73 * , 036702 ( 2006 ) .
belardinelli and v.d .
pereyra , phys .
e * 75 * , 046701 ( 2007 ) .
a.m. ferrenberg and d.p .
landau , phys .
b * 44 * , 5081 ( 1991 )
kaplan and a.n .
berker , phys .
100 * , 027204 ( 2008 ) .
kim , phys .
b * 53 * , 3388 ( 1996 ) .
m. suzuki , prog .
* 51 * , 1992 ( 1974 ) .
x. t. wu , phys . rev .
b * 79 * , 184208 ( 2009 ) .
j. w. essam , rep .
. phys . * 43 * , 33 ( 1980 ) .
x. t. wu , private cumunication .
a. malakis , a.n .
berker , i.a .
hadjiagapiou , n.g .
fytas , and t. papakonstantinou , physica a ( to appear ) , also at arxiv:0912.4586 .
a. aharony and a.b .
harris , phys .
lett . * 77 * , 3700 ( 1996 ) ; s. wiseman and e. domany , _ ibid . _ * 81 * , 22 ( 1998 ) ; s. wiseman and e. domany , phys .
e * 58 * , 2938 ( 1998 ) ; d.s .
fisher , phys .
b * 51 * , 6411 ( 1995 ) ; d.a .
huse and d.s .
fisher , _ ibid .
_ * 35 * , 6841 ( 1987 ) ; v. privman and m.e .
fisher , j. stat .
phys . * 33 * , 385 ( 1983 ) .
hartmann and h. rieger , _ optimization problems in physics _ ( wiley - vch , 2002 ) . | the effects of bond randomness on the phase diagram and critical behavior of the square lattice ferromagnetic blume - capel model are discussed .
the system is studied in both the pure and disordered versions by the same efficient two - stage wang - landau method for many values of the crystal field , restricted here in the second - order phase transition regime of the pure model . for the random - bond version
several disorder strengths are considered .
we present phase diagram points of both pure and random versions and for a particular disorder strength we locate the emergence of the enhancement of ferromagnetic order observed in an earlier study in the ex - first - order regime .
the critical properties of the pure model are contrasted and compared to those of the random model .
accepting , for the weak random version , the assumption of the double logarithmic scenario for the specific heat we attempt to estimate the range of universality between the pure and random - bond models .
the behavior of the strong disorder regime is also discussed and a rather complex and yet not fully understood behavior is observed .
it is pointed out that this complexity is related to the ground - state structure of the random - bond version . |
strong coupling qed@xmath0 has been studied for some time within the dyson - schwinger equation ( dse ) formalism both for its intrinsic interest and also as the basis for abelianized models of nonperturbative phenomena in technicolor theories and qcd . for recent reviews of dyson - schwinger equations and their application and numerous references
see for example refs .
the usual approach is to write the dse for the fermion propagator or self - energy , possibly including equations for the photon vacuum polarization or the fermion - photon proper vertex . in a recent study
@xcite it was shown for the first time how to implement nonperturbative renormalization in a numerical way within the dse formalism . in that work
the calculations were carried out in quenched approximation in landau gauge . here
we will extend these studies to arbitrary covariant gauges and will also discuss the chiral limit .
the dse s are an infinite tower of coupled integral equations and so it is always necessary to truncate this tower at some point and introduce an _ ansatz _ for any necessary undefined green s functions .
it is of course important to ensure that this _ ansatz _ be consistent with all appropriate symmetries of the theory and that it have the correct perturbative limit .
the resulting nonlinear integral equations are solved numerically in euclidean space by iteration .
dynamical ( or spontaneous ) chiral symmetry breaking ( dcsb ) occurs when the fermion propagator develops a nonzero scalar self - energy in the absence of an explicit chiral symmetry breaking ( ecsb ) fermion mass .
we will refer to coupling constants strong enough to induce dcsb as supercritical and those weaker are called subcritical .
we write the fermion propagator as @xmath1 where we refer to @xmath2 as the finite momentum - dependent fermion renormalization and where @xmath3 is the fermion mass function . in the massless theory
( i.e. , in the absence of an ecsb bare fermion mass ) by definition dcsb occurs when @xmath4 . until relatively recently
, most studies have used the bare vertex as an _ ansatz _ for the one - particle irreducible ( 1-pi ) vertex @xmath5 , @xcite despite the fact that this violates the ward - takahashi identity ( wti ) @xcite .
the resulting fermion propagator is not gauge - covariant , i.e. , physical quantities such as the critical coupling for dynamical symmetry breaking and the fermion mass pole are gauge - dependent @xcite .
there have been several studies which attempted to make the fermion dse gauge - covariant by using improved vertex forms which satisfy the wti but which possess kinematic singularities in the limit of zero photon momentum @xcite .
a general form for @xmath5 which does satisfy the ward identity and which has no unphysical kinematic singularities was given by ball and chiu in 1980 @xcite ; it consists of a minimal longitudinally constrained term which satisfies the wti , and a set of tensors spanning the subspace transverse to the photon momentum @xmath6 .
while the wti is necessary for gauge - invariance , it is not a sufficient condition and in itself does not ensure gauge covariance of the fermion propagator .
furthermore , with many vertex _ anstze _
the fermion propagator dse is not multiplicatively renormalizable , which is equivalent to saying that overlapping logarithms are present .
there has been much recent research on the use of the transverse parts of the vertex to ensure both gauge - covariant and multiplicatively renormalizable solutions @xcite , some of which will be discussed below . with the exception of ref .
@xcite , studies have mostly neglected the issue of the subtractive renormalization of the dse for the fermion propagator .
typically these studies have assumed an initially massless theory and have renormalized at the ultraviolet cutoff of the loop integration , taking @xmath7 .
where a nonzero bare mass has been used , it has simply been added to the scalar term in the propagator .
while in some circumstances for the special case of landau gauge this can be a reasonable approximation , it is in general incorrect .
although there have been earlier formal discussions of renormalization @xcite , the important step of subtractive renormalization had not been performed prior to the recent study in landau gauge @xcite . here
we present the results of a study of subtractive renormalization in the fermion dse in _ arbitrary _ covariant gauge for quenched strong - coupling qed@xmath0 .
note that here the term `` quenched '' means that the bare photon propagator is used in the fermion self - energy dse , so that @xmath8 and there is no renormalization of the electron charge .
this is a somewhat different usage to that found in lattice gauge theory studies , since in our study virtual fermion loops may still be present in the proper fermion - photon vertex .
the organization of the paper is as follows : the formalism is discussed in sec .
[ sec_formalism ] .
this section contains discussions of the dse for the renormalized fermion propagator , the _ anstze _ for the proper vertex , the subtractive renormalization procedure , the chiral limit , and renormalization point transformations .
our detailed numerical results are presented in sec .
[ sec_results ] and we present our summary and conclusions in sec . [ sec_conclusions ] .
the dse for the renormalized fermion propagator , in an arbitrary covariant gauge , is @xmath9 - i z_1(\mu,\lambda ) e^2 \int^{\lambda } \frac{d^4k}{(2\pi)^4 } \gamma^{\mu } s(k ) \gamma^{\nu}(k , p ) d_{\mu \nu}(q)\:;\ ] ] here @xmath10 is the photon momentum , @xmath11 is the renormalization point , and @xmath12 is a regularizing parameter ( taken here to be an ultraviolet momentum cutoff ) .
we write @xmath13 for the regularization - parameter dependent bare mass .
the renormalized charge is @xmath14 ( as opposed to the bare charge @xmath15 ) , and the general form for the renormalized photon propagator is @xmath16 with @xmath17 the renormalized covariant gauge parameter and @xmath18 the corresponding bare one .
since we work in the quenched approximation , we have for the coupling strength and gauge parameter respectively @xmath19 and @xmath20 , and for the photon propagator we have @xmath21 the requirement of gauge invariance in qed leads to a set of identities referred to as the ward - takahashi identities ( wti ) .
the wti for the fermion - photon vertex is @xmath22 where @xmath23 .
this is a generalization of the original differential ward identity , which expresses the effect of inserting a zero - momentum photon vertex into the fermion propagator , @xmath24 the ward identity eq .
( [ wi ] ) follows immediately from the wti of eq .
( [ wti ] ) after setting to zero all but the @xmath25 component of @xmath6 , dividing both sides of the wti by @xmath26 and then taking @xmath27 . in general , for nonvanishing photon momentum @xmath6 , only the longitudinal component of the proper vertex is constrained , i.e. , the wti provides no information on @xmath28 for @xmath29 . [ we use the notation @xmath30 and @xmath31 for the transverse and longitudinal projectors respectively . ] in particular , the wti guarantees the equality of the propagator and vertex renormalization constants , @xmath32 ( at least in any reasonable subtraction scheme @xcite . )
the wti can be shown to be satisfied order - by - order in perturbation theory and can also be derived nonperturbatively . as discussed in @xcite , this
can be thought of as just one of a set of six general requirements on the vertex : ( i ) the vertex must satisfy the wti ; ( ii ) it should contain no kinematic singularities ; ( iii ) it should transform under charge conjugation ( @xmath33 ) , parity inversion ( @xmath34 ) , and time reversal ( @xmath35 ) in the same way as the bare vertex , e.g. , @xmath36 ( where the superscript t indicates the transpose ) ; ( iv ) it should reduce to the bare vertex in the weak - coupling limit ; ( v ) it should ensure multiplicative renormalizability of the dse in eq .
( [ fermdse_eq ] ) ; ( vi ) the transverse part of the vertex should be specified to ensure gauge - covariance of the dse .
ball and chiu @xcite have given a description of the most general fermion - photon vertex that satisfies the wti ; it consists of a longitudinally - constrained ( i.e. , `` ball - chiu '' ) part @xmath37 , which is a minimal solution of the wti , and a basis set of eight transverse vectors @xmath38 , which span the hyperplane specified by @xmath39 ( i.e. , @xmath40 ) , where @xmath41 .
the minimal longitudinally constrained part of the vertex will be referred to as the ball - chiu vertex and is given by @xmath42 \gamma^\mu + \frac{(k+p)^\mu}{k^2-p^2 } \left\ { [ a(k^2 ) - a(p^2 ) ] \frac{{\not\!k}+ { \not\!p}}{2 } - [ b(k^2 ) - b(p^2 ) ] \right\}\:.\ ] ] note that since neither @xmath43 nor @xmath44 vanish identically , the ball - chiu vertex has both longitudinal and transverse components .
the transverse tensors can be conveniently written as @xcite @xmath45 ( { \not\ ! k}+{\not\ ! p})\ : , \label{t2mu}\\ t_3^\mu(k , p ) & = & q^2\gamma^{\mu}-q^{\mu}{\not \ ! q}\ : , \label{t3mu}\\ t_4^\mu(k , p ) & = & q^2\left[\gamma^{\mu}(\not\!p+\not\!k)-p^{\mu}-k^{\mu } \right ] - 2i(p - k)^{\mu}k^{\lambda}p^{\nu } \sigma_{\lambda\nu}\ : , \label{t4mu}\\ t_5^\mu(k , p ) & = & -iq_{\nu}{\sigma^{\nu\mu}}\ : , \label{t5mu}\\ t_6^\mu(k , p ) & = & \gamma^{\mu}(p^2-k^2)+(p+k)^{\mu}{\not \ !
q}\ : , \label{t6mu}\\ t_7^\mu(k , p ) & = & \frac{1}{2}(p^2-k^2)\left[\gamma^{\mu}({\not \ ! p}+ { \not \
! k})-p^{\mu}-k^{\mu}\right ] - i\left(k+p\right)^{\mu}k^{\lambda}p^{\nu } \sigma_{\lambda\nu}\ : , \label{t7mu}\\ t_8^\mu(k , p ) & = & i \gamma^{\mu}k^{\nu}p^{\lambda}{\sigma_{\nu\lambda } } + k^{\mu}{\not \ ! p}-p^{\mu}{\not \ !
k}\ : , \label{t8mu}\end{aligned}\ ] ] where we use the conventions @xmath46 , @xmath47 , and @xmath48 $ ] .
note that these tensors have been written in a different linear combination to the ones presented in ref .
a general vertex is then written as @xmath49 where the @xmath50 are functions that must be chosen to give the correct @xmath33 , @xmath34 , and @xmath35 invariance properties .
curtis and pennington published a series of articles @xcite describing their specification of a particular transverse vertex term , in an attempt to produce gauge - covariant and multiplicatively renormalizable solutions to the dse . in the framework of massless qed@xmath0
, they eliminated the four transverse vectors which are dirac - even and must generate a scalar term . by requiring that the vertex @xmath51 reduce to the leading log result for @xmath52 they were led to eliminate all the transverse basis vectors except @xmath53 , with a dynamic coefficient chosen to make the dse multiplicatively renormalizable .
this coefficient had the form @xmath54 / d(k , p)\ : , \label{cpgamma1}\ ] ] where @xmath55 is a symmetric , singularity - free function of @xmath56 and @xmath57 , with the limiting behavior @xmath58 .
[ here , @xmath2 is their @xmath59 . ] for purely massless qed , they found a suitable form , @xmath60 .
this was generalized to the case with a dynamical mass @xmath61 , to give @xmath62 ^ 2}{k^2+p^2}\ : .
\label{cpgamma2}\ ] ] they then showed that multiplicative renormalizability is retained up to next - to - leading - log order in the dcsb case .
subsequent papers established the form of the solutions for the renormalization and the mass @xcite and studied the gauge - dependence of the solutions @xcite .
dong , munczek and roberts @xcite subsequently showed that the lack of exact gauge - covariance of the solutions was due to the use of a momentum cutoff in the integral equations , since this type of regularization is not poincar invariant .
the fact that there is still some residual gauge - dependence in the physical observables such as the chiral critical point shows that with a momentum cutoff the c - p vertex _ ansatz _ is not yet the ideal choice .
dong , munczek and roberts @xcite derived an _ ansatz _ for the transverse vertex terms which satisfies the wti and makes the fermion propagator gauge - covariant under hard momentum - cutoff regularization .
bashir and pennington @xcite subsequently described two different vertex _
anstze _ which make the fermion self - energy exactly gauge - covariant , in the sense that the critical point for the chiral phase transition is independent of gauge .
specific constraints they have assumed for the vertex are , in @xcite that the transverse vertex parts vanish in the landau gauge , and in @xcite that the anomalous dimension of the fermion mass function @xmath63 is exactly 1 at the critical coupling .
their work is a continuation of that of dong , munczek and roberts , and indeed their vertex _ ansatz _ corresponds to the general form suggested in @xcite .
however , the kinematic factors @xmath64 in both vertex forms are rather complicated and depend upon a pair of as yet undetermined functions @xmath65 which must be chosen to guarantee that the weak - coupling limit of @xmath66 matches the perturbative result . renormalization studies of the dse using these new vertex _ anstze
_ should be interesting and represent a direction for further research . for the solutions to the fermion dse using the c - p vertex , the critical point for
the chiral phase transition has been shown to have a much weaker gauge - dependence than that for the dse with the bare or minimal ball - chiu vertices @xcite .
in this work we will use the curtis - pennington _ ansatz _ as the basis for our calculations .
the equations are separated into a dirac - odd part describing the finite propagator renormalization @xmath67 , and a dirac - even part for the scalar self - energy , by taking @xmath68 of the dse multiplied by @xmath69 and 1 , respectively .
the equations are solved in euclidean space and so the volume integrals @xmath70 can be separated into angle integrals and an integral @xmath71 ; the angle integrals are easy to perform analytically , yielding the two equations which will be solved numerically .
one refinement of our treatment of the c - p vertex in the present work is associated with subtleties in the ultraviolet regularization scheme .
although there have been some exploratory studies of dimensional regularization for the dse @xcite , this has not yet proven practical in nonperturbative field theory and momentum cutoffs for now remain the regularization scheme of choice in such studies .
naive imposition of a momentum cutoff destroys the gauge covariance of the dse because the self - energy integral contains terms , related to the vertex wti , which should vanish but which are nonzero when integrated under cutoff regularization @xcite . in appendix [ appdx_a ]
we derive an expression for one such undesirable term and show how it may subtracted in a simple way from the regularized self - energy .
we have also calculated some dse solutions with the usual uncorrected uv cut - off method for comparison purposes , but otherwise we use this `` gauge - improved '' regularization combined with the c - p vertex throughout this work .
this will be commented on futher in the discussion of numerical results in sec .
[ sec_results ] .
the subtractive renormalization of the fermion propagator dse proceeds similarly to the one - loop renormalization of the propagator in qed .
( this is discussed in @xcite and in @xcite , p. 425ff . )
one first determines a finite , _
regularized _ self - energy , which depends on both a regularization parameter and the renormalization point ; then one performs a subtraction at the renormalization point , in order to define the renormalization parameters @xmath72 , @xmath73 , @xmath74 which give the full ( renormalized ) theory in terms of the regularized calculation . a review of the literature of dses in qed shows , however , that this step is never actually performed .
curtis and pennington @xcite for example , define their renormalization point at the uv cutoff .
many studies take @xmath7 @xcite ; this is a reasonable approximation in landau gauge in cases where the coupling @xmath75 is sufficiently small ( i.e. , @xmath75 1 ) , but if @xmath75 is chosen large enough , the value of the dynamical mass at the renormalization point may be significantly large compared with its maximum in the infrared .
for instance , in ref .
@xcite , figures for the fermion mass are given with @xmath75 = 0.97 , 1.00 , 1.15 and 2.00 in various gauges .
for @xmath76 , the fermion mass at the cutoff is down by only an order of magnitude from its limiting value in the infrared . in general for strong coupling and/or gauges other than landau gauge
this approximation is unreliable . as shown in ref .
@xcite subtractive renormalization can be properly implemented in numerical dse studies without such approximations .
we begin with a summary of the renormalization procedure @xcite .
one defines a regularized self - energy @xmath77 , leading to the dse for the renormalized fermion propagator , @xmath78 - \sigma'(\mu,\lambda ; p ) \nonumber\\ & = & \not\!p - m(\mu ) - \widetilde{\sigma}(\mu;p ) = a(p^2)\not\!p -b(p^2)\:,\end{aligned}\ ] ] where the ( regularized ) self - energy is @xmath79 [ to avoid confusion we will follow ref .
@xcite and in this section _ only _ we will denote regularized quantities with a prime and renormalized ones with a tilde , e.g. @xmath77 is the regularized self - energy depending on both the renormalization point @xmath11 and regularization parameter @xmath12 and @xmath80 is the renormalized self - energy .
] as suggested by the notation ( i.e. , the omission of the @xmath12-dependence ) renormalized quantities must become independent of the regularization - parameter as the regularization is removed ( i.e. , as @xmath81 ) .
the self - energies are decomposed into dirac and scalar parts , @xmath82 ( and similarly for the renormalized quantity , @xmath83 ) . by imposing the renormalization boundary condition , @xmath84 one
gets the relations @xmath85 for the self - energy , @xmath86 for the renormalization constant , and @xmath87 / z_2(\mu,\lambda ) \label{baremass}\ ] ] for the bare mass .
the mass renormalization constant is then given by @xmath88 i.e. , as the ratio of the bare to renormalized mass .
the vertex renormalization , @xmath89 , is identical to @xmath90 as long as the vertex _ ansatz _ satisfies the ward identity ; this is how it is recovered for multiplication into @xmath91 in eq .
( [ reg_sigma ] ) . in order to obtain numerical solutions ,
the final minkowski - space integral equations are first rotated to euclidean space .
[ note that all equations in secs .
[ sec_intro ] and [ sec_formalism ] are written in minkowski space . ]
they are then solved by iteration on a logarithmic grid from an initial guess .
the solutions are confirmed to be independent of the initial guess and are solved with a wide range of cutoffs ( @xmath12 ) , renormalization points ( @xmath11 ) , couplings ( @xmath75 ) , covariant gauge choices ( @xmath17 ) , and renormalized masses ( @xmath92 ) .
the chiral limit occurs by definition when the bare mass is taken to zero sufficiently rapidly as the regularization is removed .
this is guaranteed , for example , by maintaining @xmath93 as @xmath94 .
explicit chiral symmetry breaking ( ecsb ) occurs when the bare mass @xmath13 is not zero , ( or more precisely , whenever it is not taken to zero sufficiently rapidly as @xmath95 ) .
dynamical chiral symmetry breaking ( dcsb ) is said to have occured when @xmath96 in the absence of ecsb .
as the coupling strength increases from zero , there is a transition to a dcsb phase at the critical coupling @xmath97 .
concisely , the absence of ecsb means that we can set @xmath93 , and the absence of both ecsb and dcsb ( i.e. , @xmath98 ) means that @xmath61 , @xmath92 , and @xmath13 simultaneously vanish .
( recall that in the notation that we use here , @xmath99 and @xmath100 . )
this is the same definition of the chiral limit that is used in nonperturbative studies of qcd , see e.g. refs .
@xcite and references therein .
obviously , any limiting procedure where we take @xmath101 sufficiently rapidly as @xmath94 will also lead to the chiral limit @xcite .
a renormalization point transformation is a change of renormalization scale [ i.e. , @xmath102 for @xmath103 such that the bare mass(es ) and coupling(s ) remain fixed for fixed regularization parameter ( @xmath12 ) and fixed renormalization scheme .
this ensures that the physical observables of the theory are invariant under such a transformation .
this set of transformations is associative [ @xmath104 , contains the identity [ @xmath105 , and contains all inverses [ @xmath106 and hence is called the renormalization group . for the purposes of the discussion here we will now explicitly indicate the choice of renormalization point by a @xmath11-dependence of the renormalized quantities , i.e. , @xmath107 , @xmath108 , etc . note that eq .
( [ ren_inv_s ] ) implies that @xmath109 the renormalization point boundary condition in eq .
( [ ren_point_bc ] ) then leads to @xmath110 , or equivalently , to the two boundary conditions @xmath111 and @xmath112 . from eq .
( [ ab_at_mu ] ) and the fact that @xmath113 we have @xmath114&= & 1-\left[\sigma'_d(\mu,\lambda;p^2)/z_1(\mu,\lambda)\right ] \ ; , \nonumber\\ \left[b(\mu;p^2)/z_2(\mu,\lambda)\right]&=&m_0(\lambda)+ \left[\sigma'_s(\mu,\lambda;p^2)/z_1(\mu,\lambda)\right]\;. \label{ab_rescaled}\end{aligned}\ ] ] consider the effects of an an arbitrary rescaling @xmath115 and @xmath116 , [ i.e. , @xmath61 fixed ] , for some real constant @xmath117 .
it is straightforward to see that under such a rescaling we have @xmath118 and @xmath119 .
it follows that the rh sides of eqs .
( [ ab_rescaled ] ) are unaffected by such an arbitrary rescaling .
hence , it follows that the choice of renormalization point boundary conditions is equivalent to the choice of scale for the functions @xmath120 and @xmath121 .
let us consider this observation in more detail .
since we are working in the quenched approximation , where @xmath122 and @xmath123 are unaffected by a change of renormalization point , it follows from eq .
( [ reg_sigma ] ) that @xmath124 is renormalization point independent since a change of renormalization point is a rescaling of @xmath120 and @xmath121
. then since @xmath13 is renormalization point independent by definition , the entire rhs of eqs .
( [ ab_rescaled ] ) must be independent of the choice of renormalization point .
thus , under a renormalization point transformation we must have _ for all _
@xmath125 @xmath126 from which it follows for the fermion propagator that @xmath127 in the usual way .
the behavior in eq .
( [ ren_pt_transf ] ) is explicitly tested for our numerical solutions .
it is clear from eq .
( [ ren_pt_transf ] ) that having a solution at one renormalization point ( @xmath11 ) completely determines the solution at any other renormalization point ( @xmath128 ) without the need for any further computation .
an alternative derivation of this result which starts from the renormalized action and which applies to the general unquenched case can be found for example in sec .
2.1 of ref .
@xcite . for brevity we can denote the above renormalization point dependence of the fermion propagator by @xmath129 . in the general unquenched case
@xcite we would have in addition @xmath130 , @xmath131 , and @xmath132 .
solutions were obtained in euclidean space for the dse for couplings @xmath75 from 0.1 to 1.30 , in gauges with @xmath17 from -0.25 to 3 , and with a variety of renormalization points and renormalized masses . all results in this section
refer to euclidean space quantities . in the graphs and tables that follow ,
there are no explicit mass units .
since the equations have no inherent mass - scale , the cutoff @xmath12 , renormalization point @xmath11 , @xmath92 , and units of @xmath61 or @xmath133 all scale multiplicatively , and the units are arbitrary . in four dimensions the coupling has no mass dimension , therefore it remains unchanged for all such choices of mass units .
[ diff_gauges ] shows a family of solutions characterized by @xmath134 , @xmath135 , @xmath136 , and gauge parameters from -0.25 to 1.25 .
we see that while @xmath120 and @xmath121 are strongly gauge dependent , the mass function @xmath3 is relatively insensitive to @xmath17 .
the location of the mass pole of the physical electron must of course be independent of gauge , and this gauge independence has been demonstrated explicitly using the wti for example by atkinson and fry @xcite .
their proof assumes that the bare mass @xmath13 is itself independent of gauge .
hence , in a fully gauge covariant treatment the mass function is independent of gauge at two scales ( i.e. , at the mass pole and at the uv regularization scale @xmath12 ) . in our study
we find that the mass function is relatively insensitive to the choice of gauge for all @xmath125 .
the nature of the landau - khalatnikov transformations @xcite makes the possibility of a gauge independent @xmath61 seem rather unlikely .
the stability of the renormalized dse solutions with respect to variations in the ultraviolet cutoff is evident in fig .
[ diff_uv ] .
this graph shows solutions with @xmath137 , @xmath138 , @xmath136 , and gauge @xmath139 .
the cutoff @xmath140 was varied over several orders of magnitude with no apparent change in the solutions over the common range of momenta .
this numerical stability was shown in other tests as well .
for instance , we extracted the mass @xmath141 for solutions with @xmath134 , @xmath142 , @xmath143 , and observed variations of less than one part in @xmath144 as the uv cutoff was varied over six orders of magnitude .
tables [ landau_tbl ] , [ gp025_tbl ] , and [ gp050_tbl ] show the evolution of the renormalization constants @xmath90 , @xmath145 , and the cutoff - dependent bare mass @xmath13 as a function of the uv regulator @xmath12 .
we see that as we move further from landau gauge @xmath90 decreases more rapidly with increasing @xmath12 .
in addition , we observe that the bare mass exhibits decaying oscillations with increasing @xmath12 , which is directly related to the oscillations characteristic of the supercritical coupling and subtractive renormalization at large @xmath125 ( see additional discussion later ) . fig
. [ diff_alpha ] shows solutions with the coupling varying from subcritical ( @xmath146 ) to supercritical ( @xmath147 ) values , with identical renormalization point , renormalized mass and gauge .
here we see that the the nodes in the mass function @xmath61 move to lower momenta and the oscillations become more pronounced as the coupling is increased further and further above critical coupling . in order to test the gauge - invariance of the chiral critical point , we extracted the critical coupling from solutions in the landau gauge and in two other covariant gauges , with @xmath148 and 0.5 .
_ , @xcite working in the quenched ladder approximation in landau gauge , found that the infrared limit of the dynamical mass @xmath149 has an infinite - order phase transition , @xmath150\:.\ ] ] following their approach , we assume a similar form for the dynamical mass near the critical coupling , @xmath151 \:,\ ] ] and construct an order parameter which is expected to have a second - order phase transition , @xmath152 $ ] . since the inherent mass scale @xmath153 is not known _ a priori _ , it is necessary to choose a reasonable scale @xmath154 , and then calculate a corrected fit , which also yields the actual value of @xmath153 .
dse solutions were obtained for several values of the coupling in each gauge , with the renormalization @xmath155 .
we set @xmath156 for the purpose of studying the transition .
the ir mass limit was extrapolated for each solution and then the order parameters were calculated using an assumed mass scale @xmath157 .
the resulting critical curves are shown along with the @xmath149 values from the solutions , in fig .
[ crit_curves ] .
the parameters from the nonlinear fits are given in table [ crit_fits_tbl ] .
the critical exponents @xmath158 are the same to within their numerical tolerance , and suggest that @xmath158 may be independent of gauge .
although the values of @xmath97 are close in value , there is clear evidence of residual gauge - dependence .
our value for @xmath97 in the landau gauge is very close to the value of 0.933667 found by atkinson _
et al _ @xcite in their bifurcation analysis of the solutions of the fermion dse with the curtis - pennington vertex .
in addition we find that the critical coupling varies with @xmath17 in the same direction .
recall that these authors used the unrenormalized equations and relaxed the uv momentum cutoff to infinity in order to remove cutoff artifacts , whereas we have used subtractive renormalization and our gauge covariance correction .
hence we can anticipate a small difference between the critical couplings in our approaches .
[ diff_mu ] shows a family of equivalent solutions renormalized at different momentum scales @xmath11 and these provide a direct check on the behavior predicted in eq .
( [ ren_pt_transf ] ) .
all have coupling @xmath159 and @xmath160 , and the renormalization scale @xmath161 is stepped by powers of 10 , with renormalized masses @xmath92 chosen such that @xmath162 for each @xmath11 .
it is clear that the resulting mass curves are identical at all @xmath125 as expected , and in each case @xmath163 scales as predicted in eq .
( [ ren_pt_transf ] ) showing that the renormalized dse does transform correctly .
the anomalous dimension of the mass , @xmath63 , is defined by the asymptotic scaling of the dynamical mass with @xmath125 , @xmath164 as well as depending on the coupling , it shows a slight dependence on the gauge , as shown in fig .
[ scaled_mass ] .
the dynamical masses shown are from dse solutions with @xmath165 , @xmath138 , and @xmath136 , in landau gauge and in gauges @xmath139 and 0.5 .
they are scaled by multiplication with @xmath166 , where the value of @xmath63 used was that extracted from the landau gauge solution , @xmath167 .
the gauge - dependence of @xmath63 shows up as the slight difference in slopes on the log - log plot .
( the dips apparent at the end of the curves are due to having a hard momentum cut - off @xmath168 . as @xmath12 is increased these move to higher momenta also . ) in landau gauge , the actual power of @xmath169 with which @xmath61 falls asymptotically is @xmath170 . for the gauges @xmath139 and 0.5 , the anomalous dimensions are 1.713948 and 1.711274 , giving powers of @xmath169 equal to 0.143026 and 0.144363 respectively .
miransky has studied the form of the mass renormalization @xmath171 in the bare vertex approximation in landau gauge , and without subtractive renormalization @xcite . in this treatment
he finds @xmath172 , with the exponent @xmath173 where the critical coupling for dcsb in that approximation is @xmath174 .
this would imply an asymptotic scaling for the dynamical mass that goes like @xmath175 , so that the anomalous mass dimension would be related to @xmath176 by @xmath177 .
recent articles by holdom @xcite and mahanta @xcite claim that for quenched theories _ at criticality _ the mass anomalous dimension @xmath63 should be exactly 1 , giving @xmath178 as in the bare approximation , and that in particular this result should be independent of the gauge @xcite . in miransky s treatment
this corresponds to the vanishing of @xmath176 at the critical coupling .
we find for subcritical couplings that @xmath179 and that it is gauge insensitive .
hence , as for the pure cp - vertex case , our gauge corrected cp vertex does not lead to @xmath180 .
the gauge covariance correction described in the appendix leads to an exact restoration of gauge covariance in the subcritical case with no explicit chiral symmetry breaking , i.e. , no bare mass ( see , e.g. , ref .
the difference between a dse solution with naive cutoff regularization and those with the gauge covariance correction is shown in fig .
[ gauge_cov_fix ] .
both solutions have @xmath159 , @xmath160 , and are renormalized at @xmath138 and @xmath136 . the quantitative change induced by our gauge covariance correction
was found to be relatively small in the presence of a substantial mass function @xmath61 .
we find , as in our previous study in landau gauge @xcite , that for supercritical couplings the dynamic mass crosses zero ; for solutions away from the chiral limit , the position of the first node depends on the gauge , as shown in fig .
[ mass_nodes ] .
in fact as the cutoff is increased @xmath61 shows damped oscillations periodic in @xmath181 , as shown in fig .
[ mass_damped_osc ] .
this has been discussed by several authors @xcite ; in particular it is shown using some simplifying approximations , in @xcite that in landau gauge , the dse reduces to a differential equation for @xmath182 which has the solution @xmath183 with @xmath184
. however , the approximations used in deriving this result are not applicable outside landau gauge and even in landau gauge lead to differences from the present treatment .
we find that the functional form is substantially correct , but that both the mass dimension and the period of oscillations depend on the coupling and the choice of gauge .
in fact , for the case shown in fig .
[ mass_damped_osc ] , the mass dimension that fits @xmath185 is @xmath186 .
the dependence of the period on @xmath75 and @xmath97 is also not as simple as that in eq .
( [ mass_osc_form ] ) .
we have extended our previous work on the numerical renormalization of the dse @xcite to arbitrary covariant gauges .
the procedure is straightforward to implement and extremely stable .
it becomes numerically more challenging for covariant gauges far removed from landau gauge and for large couplings ( @xmath187 ) .
the importance of the approach is that it removes the issue of cut - off dependence and allows solutions to be obtained for any choice of renormalization point .
we have described the procedure for performing renormalization group transformations between solutions with different renormalization points .
we saw that a knowledge of the solution at one renormalization point automatically provides the solutions at all renormalization points .
this then allows also comparisons with results from lattice studies of qed , which should prove useful in providing further guidance in the choice of reasonable _ anstze _ for the vertex and photon propagator . without renormalization only the unrenormalized , regulated quantities
would be be obtained and any such comparisons would be meaningless .
in addition , in order to study the nonperturbative behaviour of renormalization constants such as @xmath89 , @xmath90 , and @xmath145 they must be numerically extracted and so a method such at that described here would be essential . the context of this study has been quenched four - dimensional qed with a modified curtis - pennington vertex , since that vertex _ ansatz _ has the desirable properties of making the solutions approximately gauge - invariant and also multiplicatively renormalizable up to next - to - leading log order .
the technique described can be generalized to apply elsewhere ( e.g. , qcd ) , whenever numerical renormalization is required .
the solutions are stable and the renormalized quantities become independent of regularization as the regularization is removed , which is as expected . for example , the mass function @xmath61 and the momentum - dependent renormalization @xmath2 are unchanged to within the numerical accuracy of the computation as the integration cutoff is increased by many orders of magnitude .
the mass renormalization constant @xmath145 converges to zero with increasing @xmath12 because the mass function @xmath61 falls to zero sufficiently rapidly at large @xmath125 .
the absence of divergences of @xmath113 , @xmath13 , and @xmath145 in the limit @xmath81 is a purely nonperturbative result and is in sharp contrast to the perturbative case where these constants diverge at all orders . in order to study the critical point and exponents for the transition to dcsb ,
one sets the renormalized mass @xmath92 to 0 and varies the coupling . for subcritical couplings
@xmath13 remains zero , and the dynamical mass @xmath61 is identically zero , while for supercritical couplings @xmath4 [ and @xmath188 for finite @xmath12 ] .
we have extracted the critical coupling for dcsb in landau gauge ( @xmath190 ) , and in gauges with @xmath139 and 0.5 .
our landau gauge result is very close to the value found in @xcite ; the values in the other gauges show a small residual gauge - dependence . for subcritical couplings , we find that the mass renormalization @xmath145 scales approximately as @xmath191 , where e.g. @xmath192 . above the critical coupling ,
the mass function shows damped oscillations around zero , periodic in @xmath193 .
we have extracted mass anomalous dimensions @xmath63 for some subcritical and supercritical couplings , and find them all greater than 1 .
we have shown that our modified curtis - pennington vertex , while removing the violation of gauge covariance in the massless subcritical case ( i.e. , when there is no explicit or dynamical chiral symmetry breaking ) , has not been sufficient to remove the small residual violation of gauge covariance in the general case .
hence it is important to attempt to extend this work to include other regularization schemes ( e.g. , dimensional regularization ) and to vertices of the bashir - pennington type @xcite .
we thank michael neuling for assistance with some of the numerical calculations .
this work was partially supported by the australian research council , by the u.s .
department of energy through contract no .
de - fg05 - 86er40273 , and by the florida state university supercomputer computations research institute which is partially funded by the department of energy through contract no .
de - fc05 - 85er250000 .
the work of cdr was supported by the us department of energy , nuclear physics division , under contract number w-31 - 109-eng-38 .
this research was also partly supported by grants of supercomputer time from the u.s . national energy research supercomputer center and the australian national university supercomputer facility .
agw thanks the ( department of energy ) institute for nuclear theory at the university of washington for its hospitality and partial support during the completion of this work .
regulators applied to divergent integrals in field theory always destroy some continuous symmetry , and in particular the use of a momentum cutoff destroys gauge covariance .
this appendix describes a modification of the self - energy integrals in the regularized dse , which will at least partially restore this symmetry .
the basis of this change in the regularization scheme is that when the self - energy @xmath194 , given in eq .
( [ reg_sigma ] ) , is evaluated under cutoff regularization , it contains a term related to the vertex wti , which should vanish but which integrates to give a nonzero contribution because the cutoff regularization scheme is not translationally invariant @xcite .
we therefore evaluate this term separately and subtract it from the self - energy .
it turns out to affect only the value of the renormalization @xmath90 .
we write the photon propagator , in minkowski momentum space and in an arbitrary covariant gauge , as @xmath195 where @xmath196 and @xmath197 are the transverse and longitudinal projectors respectively , @xmath198 , @xmath199 .
the renormalized fermion propagator is as in eq .
( [ fermprop_formal ] ) .
@xmath200 is the renormalized proper vertex ; @xmath23 is the photon momentum . the renormalization point for @xmath201 , @xmath202 , and @xmath203
is @xmath204 , however , we will not always write it explicitly . the `` naively '' regularized self - energy ( under a regularization scheme with parameter @xmath12 ) is @xmath205 ( where we write the regularization as though it were a momentum cutoff , but it need not be ) .
if @xmath5 satisfies the wti , @xmath206 we can rewrite the dse as @xmath207 the boxed integral is odd in @xmath6 , and should vanish in any translationally invariant regularization scheme ; otherwise , it contributes and we expect it to destroy the gauge - covariance of @xmath208 .
however , we can define a `` gauge - improved '' self - energy by cancelling this undesirable term @xmath209 and since the added integral is dirac - odd , upon decomposing into scalar and spinor parts as in eq .
( [ decompose ] ) we have @xmath210 this modification combined with the c - p vertex is the _ ansatz _ used in all calculations in this work unless explicitly stated otherwise . converting to euclidean metric ( but suppressing the `` euclidean '' subscript on momenta for convenience ) gives @xmath211 since in our case
the regularization is an ultraviolet cutoff in momentum , we have @xmath212 introducing the variables @xmath213 , @xmath214 , @xmath215 , @xmath216 , @xmath217 , we have @xmath218 thus the gauge covariance correction that we use here is to cancel the boxed term in eq .
( [ eq_box ] ) by adding @xmath219 to the naive regularized self - energy in eq .
( [ eq_naive ] ) and this is how it is implemented in our program .
the effects of including and not including this gauge covariance correction can be studied numerically .
roberts and a.g .
williams , _ dyson - schwinger equations and their application to hadronic physics _ , in _ progress in particle and nuclear physics , vol . 33 _ ( pergamon press , oxford , 1994 ) , p. 477 .
v. a. miranskii , _ dynamical symmetry breaking in quantum field theories _ , ( world scientific , singapore , 1993 ) .
i. fomin , v. p. gusynin , v. a. miransky and yu .
a. sitenko , riv .
nuovo cim .
* 6 * , 1 ( 1983 ) . f.t .
hawes and a.g .
williams , phys .
d * 51 * , 3081 ( 1995 ) .
r. haag and th.a .
maris , phys .
rev . * 132 * , 2325 ( 1963 ) ; th.a.j .
maris , v.e .
herscovitz and d. jacob , phys .
* 12 * , 313 ( 1964 ) ; t. nonoyama and m. tanabashi , prog .
phys . * 81 * , 209 ( 1989 ) .
i. fomin , v. p. gusynin , and v. a. miransky , phys .
* 78b * , 136 ( 1978 ) . v. a. miransky , sov .
jetp * 61 * ( 5 ) , 905 ( 1985 ) .
v. a. miransky , nuovo cimento * 90a * , 149 ( 1985 ) .
e. l. rakow , nucl .
b356 * , 27 ( 1991 ) .
ward , phys .
rev . * 78 * , 124 ( 1950 ) ; y. takahashi , nuovo cimento * 6 * , 370 ( 1957 ) .
aoki , m. bando , t. kugo , k. hasebe and h. nakatani , prog .
* 81 * , 866 ( 1989 ) .
curtis and m.r .
pennington , phys .
* , 4933 ( 1993 ) .
kondo , y. kikukawa and h. mino , phys .
* 220b * , 270 ( 1989 ) .
d. atkinson , p. w. johnson and k. stam , phys .
* 201b * , 105 ( 1988 ) .
j.s . ball and t.w .
chiu , phys .
rev . d*22 * , 2542 ( 1980 ) ; _ ibid .
_ , 2550 ( 1980 ) .
king , phys .
rev . d*27 * , 1821 ( 1983 ) .
p. rembiesa , phys .
d*41 * , 2009 ( 1990 ) .
b. haeri , phys .
rev . d*43 * , 2701 ( 1991 ) .
curtis and m.r .
pennington , phys .
* , 4165 ( 1990 ) .
curtis and m.r .
pennington , phys .
* , 536 ( 1991 ) .
curtis and m.r .
pennington , phys .
d*46 * , 2663 ( 1992 ) . z. dong , h. munczek , and c.d .
roberts , phys .
333b * , 536 ( 1994 ) .
a. bashir and m. r. pennington , phys .
d * 50 * , 7679 ( 1994 ) .
a. bashir and m. r. pennington , phys .
d * 53 * , 4694 ( 1996 ) .
roberts , _ schwinger dyson equations : dynamical chiral symmetry breaking and confinement _ , in _ qcd vacuum structure _ , edited by h. m. fried and b. mller ( world scientific , singapore , 1993 ) .
a. kizilers , m. reenders , and m. r. pennington , phys .
d * 52 * , 1242 ( 1995 ) .
d. atkinson , j.c.r .
bloch , v.p .
gusynin , m. r. pennington , and m. reenders , phys . lett .
* 329b * , 117 ( 1994 ) .
l. von smekal , p. a. amundsen , and r. alkofer , nucl .
* a529 * , 633 ( 1991 ) ; m. becker , `` nichtperturbative strukturuntersuchungen der qed mittels genherter schwinger - dyson - gleichungen in dimensioneller regularisierung , '' , ph.d .
dissertation , w. w. u. mnster,1995 . c. itzykson and j.b .
zuber , _ quantum field theory _ , ( mcgraw - hill , new york , 1980 ) .
h. j. rothe , _ latticd gauge theories : an introduction _ , ( world scientific , singapore , 1992 ) .
d. atkinson and m.p .
fry , nucl .
b156 , 301 ( 1979 ) .
l. d. landau and i. m. khalatnikov , sov .
jetp * 2 * , 69 ( 1956 ) [ translation of zhur .
* 29 * , 89 ( 1955 ) ] ; k. johnson and b. zumino , phys . rev .
lett . * 3 * , 351 ( 1959 ) .
b. holdom , phys .
lett . * 62 * , 997 ( 1989 ) ; ( r ) _ ibid . _ * 63 * , 1889 ( 1989 ) ; u. mahanta , phys .
* 225b * , 181 ( 1989 ) .
u. mahanta , phys .
lett . * 62 * , 2349 ( 1989 ) .
m. r. frank and c. d. roberts , phys .
d * 53 * , 390 ( 1996 ) . .
renormalization constant @xmath90 , bare masses @xmath13 , and mass renormalization @xmath145 , as a function of uv cutoff for @xmath137 in the landau gauge ( @xmath220 )
. all solutions are with renormalization point @xmath221 and renormalized mass @xmath222 [ cols= " < , > , < , > , < , > , < " , ] | we extend a previous landau - gauge study of subtractive renormalization of the fermion propagator dyson - schwinger equation ( dse ) in strong - coupling , quenched qed@xmath0 to _ arbitrary _ covariant gauges .
we use the fermion - photon proper vertex proposed by curtis and pennington with an additional correction term included to compensate for the small gauge - dependence induced by the ultraviolet regulator .
we discuss the chiral limit and the onset of dynamical chiral symmetry breaking in the presence of nonperturbative renormalization .
we extract the critical coupling in several different gauges and find evidence of a small residual gauge - dependence in this quantity . |
the modelling and simulation of pedestrians and crowds is a consolidated and successful application of research results in the more general area of computer simulation of complex systems .
it is an intrinsically interdisciplinary effort , with relevant contributions from disciplines ranging from physics and applied mathematics to computer science , often influenced by ( and sometimes in collaboration with ) anthropological , psychological , sociological studies and the humanities in general .
the level of maturity of these approaches was sufficient to lead to the design and development of commercial software packages , offering useful and advanced functionalities to the end user ( e.g. cad integration , cad - like functionalities , advanced visualisation and analysis tools ) in addition to a simulation engine . nonetheless , as testified by a recent survey of the field by @xcite and by a report commissioned by the cabinet office by @xcite , there is still much room for innovations in models improving their performances both in terms of _ effectiveness _ in modelling pedestrians and crowd phenomena , in terms of _ expressiveness _ of the models ( i.e. simplifying the modelling activity or introducing the possibility of representing phenomena that were still not considered by existing approaches ) , and in terms of _ efficiency _ of the simulation tools .
in addition to the above directions , we want to emphasise the fact that one of the sometimes overlooked aspects of a proper simulation project is related to the _ calibration _ and _ validation _ of the results of tools related to the _ synthesis _ of the pedestrians and crowd behaviour in the considered scenario .
these phases are essentially related to the availability of proper empirical data about or , at least , relevant to , the considered scenario ranging from the pedestrian demand ( i.e. an origin destination matrix ) , preferences among different alternative movement choices ( e.g. percentage of persons employing stairs , escalators and elevators in a multiple level scenario ) , but also the average waiting times at service points ( i.e. queues ) , the average time required to cover certain paths , the spatial distribution of pedestrians in specific environmental conditions that is required to evaluate the so called `` level of service '' associated to portions of the environment as defined by @xcite .
these data are results of activities of _ analysis _ , some of which can be fruitfully automated given , on one hand , the wide diffusion of cameras employed for video surveillance of public areas and , on the other , considering the level of maturity of video processing and analysis techniques
. an _ integrated approach _ to pedestrians and crowd studies encompasses both the application of analysis and synthesis techniques that , in a virtuous circle , can mutually benefit one from the other , to effectively ( i ) identify , ( ii ) face and ( iii ) provide innovative solutions to challenges towards the improvement of the understanding of crowding phenomena .
this kind of interdisciplinary study , in addition to computer science , often employs or directly involves research results in the area of social sciences in general , can be found in the literature .
for instance , @xcite show how computational fields guiding simulated pedestrians movement in a simulated environment can be automatically derived by video footages of actual people moving in the same space .
@xcite , instead , employ an hydrodynamic model , that ( to a certain extent ) can represent the flow of pedestrians in mass movement situations , to improve the characterisation of pedestrian flows by means of automatic video analysis . from the perspective of offering a useful service to crowd managers @xcite
describe an anticipative system integrating computer vision techniques and pedestrian simulation able to suggest crowd management solutions ( e.g. guidance signals ) to avoid congestion situations in evacuation processes .
finally , in @xcite the authors propose the employment of the _ social force model _ by @xcite , probably the most successful example of crowd synthesis model , to support the detection of abnormal crowd behaviour in video sequences .
an example of identification of a still not considered phenomenology is related to a work by @xcite in which the authors have defined an extension to the social force model by that considers the presence of groups in the simulated population : the motivations and some modelling choices ( i.e. the limited size of considered groups and their spatial arrangement ) are based on actual observations and analyses .
a related effort , carried out instead by a research group trying to improve crowd analysis techniques , is described in @xcite : in this case , the social force model acts as a sort of predictor block in an automated video analysis pipeline , improving the tracking in case of groups within the observed flow .
finally , @xcite also focus on groups , as central element of an observation and analysis that also considers psychometric factors .
it is important to emphasise that anthropological considerations about human behaviour in crowded environments , such as the analysis of spatial social interaction among people , are growingly considered as crucial both in the computerised analysis of crowds as pointed out by @xcite and in the synthesis of believable pedestrian and crowd behaviour , such as in @xcite and in @xcite .
in particular , _ proxemics _ has a prominent role both in the modelling and analysis of pedestrians and crowd behaviour .
the term was first introduced by @xcite with respect to the study of a type of non verbal behaviour related to the dynamic regulation of interpersonal distance between people as they interact . within this context the aim of this paper
is to provide a comprehensive framework comprising both the synthesis and analysis of pedestrians and crowd behaviour : in this schema we suggest both ways in which the results of the analysis can provide fruitful inputs to modellers and , on the other hand , how results of the modelling and simulation activities can contribute to the ( automated ) interpretation of raw empirical data .
the framework will be described in the following section , while an example of unfolding of these conceptual and experimental pathways will be described through the introduction of an adaptive model for group cohesion ( section [ sec : model ] ) that was motivated and that will effectively be calibrated and validated by means of analyses on observed crowd behaviour ( section [ sec : observation ] ) .
conclusions and future works will end the paper .
a comprehensive framework trying to put together different aspects and aims of pedestrians and crowd dynamics research has been defined in @xcite .
the central element of this schema is the mutually influencing ( and possibly motivating ) relationship between the above mentioned efforts aimed at synthesising crowd behaviour and other approaches that are instead aimed at analysing field data about pedestrians and crowds in order to characterise these phenomena in different ways .
it must be noted , in fact , that some approaches have the goal of producing aggregate level quantities ( e.g. people counting , density estimation ) , while others are aimed at producing finer - grained results ( i.e. tracking people in scenes ) and some are even aimed at identifying some specific behaviour in the scene ( e.g. main directions , velocities , unusual events ) .
the different approaches adopt different techniques , some performing a _ pixel level analysis _
, others considering larger patches of the image , i.e. _ texture level analysis _ ; other techniques require instead the detection of proper objects in the scene , a real _ object
level analysis_. from the perspective of the requirements for the synthesis of quantitatively realistic pedestrian and crowd behaviour , it must be stressed that both aggregate level quantities and granular data are of general interest : a very important way to characterise a simulated scenario is represented by the so called fundamental diagram by @xcite , that represents the relationship in a given section of an environment between the flow of pedestrians and their density .
qualitatively , a good model should be able to reproduce an empirically observed phenomenon characterised by the growth of the flow until a certain density value ( also said _
critical density _ ) is reached ; then the flow should decrease .
however , every specific situation is characterised by a different shape of this curve , the position of critical density point and the maximum flow level ; therefore even relatively `` basic '' counting and density estimation techniques can provide useful information in case of observations in real world scenarios .
density estimation approaches can also help in evaluating qualitatively the patterns of space utilisation generated by simulation models against real data .
tracking techniques instead can be adopted to support the estimation of traveling times ( and length of the followed path ) by pedestrians .
crowd behaviour understanding techniques can help in determining main directions and the related velocities . to complete the initial schema , suggesting that analysis and synthesis should mutually benefit one from the other , we propose here the extension of a different kind of diagram , that can be found in @xcite or more recently in @xcite , used to discuss the form of inference that can be carried out by using a model as a method of study .
starting from portion of the reality that we will call _ target system _ the process of synthesis leads to the definition of a model and its implementation in the form of a simulator .
the latter can then be employed ( i.e. executed with specific inputs and parameters ) to carry out a simulation campaign leading to a set of results .
the processes of analysis involve , on one hand , raw data that can be acquired through direct observations and controlled experiments . on the other hand ,
also simulation results require a process of interpretation in order to be comparable with observed empirical data .
when this cycle produces simulation results that , once interpreted and analysed , actually match the empirical data acquired on the field , the defined model can be employed for sake of explanation and prediction .
of course it is not immediate to define a model that generates simulation results matching empirical data , especially since models of complex systems are generally characterised by a number of parameters even when the modeller tries to keep the model as simple and elegant as possible .
it is this need of actually _ calibrating _ model parameters for achieving a model _ validation _ , as described by @xcite , that actually introduces the first type of synergetic collaboration between analysis and synthesis : the analysis of raw data about the simulated phenomenon leads to the possibility of identifying values or at least intervals for model parameters .
this is not actually the only case of influence of the analysis on the definition of models : in fact , it is the observation of the system that leads to the identification of phenomenologies that are not currently represented and managed by a model and that can represent the motivations and goals for model innovation .
the potential outcomes of the modelling and simulation phases that can have an influence on the analysis activities are related to two categories of contributions : on one hand , the need to create a mechanism for the generation of an observed phenomenon leads to its formalisation , that could be instrumental also in the creation of additional mechanisms for its automated analysis .
on the other hand , even long before reaching the necessary level of maturity of the simulator , the modeller and developer of a simulation system actually needs to define and develop metrics , indicators and techniques to evaluate the outcomes of the modelling phases .
these by - products of the synthesis activity can also represent a starting point for the actual development of automated analysis approaches .
the following sections actually represent an unfolding of this kind of conceptual and experimental process .
in particular , section [ sec : model ] introduces a model for pedestrian and group behaviour representing the evolution of a first approach that was started to face issues raised by an unstructured and non - systematic observation of crowd patterns and movements in a real world scenario described by @xcite . in this line of research ,
a model encompassing groups as a fundamental element influencing the overall system dynamics was designed and implemented and a metric for the evaluation of group cohesion ( and , therefore , also its dispersion ) was defined .
this metric led to the understanding that the first version of the model was unable to preserve the cohesion of groups and therefore it was instrumental in the realisation of a new model described in @xcite that endows pedestrians that are also members of specific types of group with an adaptive mechanism to preserve their cohesion even in difficult situations ( e.g. presence of obstacles or high density environments ) .
this research effort also led to the identification of new observations and analyses to back - up simulation results with empirical data ; the dispersion metric was also the starting point for the formal definition of measurements to be executed in the analysis of the acquired raw data .
these activities will be described in section [ sec : observation ] .
this section introduces a model representing pedestrian behaviour in an environment , considering the impact of the presence of simple and structured groups .
the model is characterised by a discrete representation of the environment and time evolution , and it is based on the floor - field mechanism of existing ca approaches .
however , the pedestrian behaviour is so articulated , comprising an adaptive mechanism for the preservation of group cohesion , to the point that the model is more properly classified as agent - based .
the different elements of the model will now be introduced , then some results of its application to sample simulation scenarios will be given to show the model capabilities and the requirements in terms of empirical data to complete the calibration of the group cohesion adaptive mechanism .
a more detailed formal introduction of the model and additional simulation results can be found in @xcite .
the physical environment is represented in terms of a discrete grid of square cells : @xmath0 where @xmath1 .
the size of every cell is @xmath2 according to standard measure used in the literature and derived from empirical observation and experimental procedure shown in @xcite and @xcite .
every cell has a row and a column index , which indicates its position in the grid : @xmath3 .
consequently , a cell is also identified by its row and column on the grid , with the following notation : @xmath4 .
every cell is linked to other cells , that are considered its neighbours according to the moore neighbourhood .
cells are annotated and virtual grids are superimposed on the base environmental representation to endow the environment with the capability to host pedestrian agents and support their perception and action .
markers are sets of cells that play particular roles in the simulation
. three kinds of marker are defined in the model : ( i ) _ start areas _ , places ( sets of cells ) were pedestrians are generated ; they are characterised by information for pedestrian generation both related to the type of pedestrians and to the frequency of generation . in particular
, a start area can generate different kinds of pedestrians according to two approaches : ( a ) _ frequency - based generation _
, in which pedestrians are generated during all the simulation according to a frequency distribution ; ( b ) _ en - bloc generation _ , in which a set of pedestrians is generated at once in the start area when the simulation starts ; ( ii ) _ destination areas _ , final places where pedestrians might want to go ; ( iii ) _ obstacles _ , non - walkable cells defining obstacles and non - accessible areas . adopting the approach of the _ floor field
_ model by @xcite , the environment of the basic model is composed also of a set of superimposed virtual grids , structurally identical to the environment grid , that contains different floor fields that influence pedestrian behaviour . the goal of these grids is to support long range interactions by representing the state of the environment ( namely , the presence of pedestrians and their capability to be perceived from nearby cells ) . in this way ,
a local perception for pedestrians actually simply consists in gathering the necessary information in the relevant cells of the floor field grids .
floor fields are either _ static _ ( created at the beginning and not changing during the simulation ) or _ dynamic _ ( changing during the simulation ) .
three floor fields are considered in the model : * the _ path field _ assigned to each destination area : this field indicates for every cell the distance from the destination , and it acts thus as a gradient , a sort of potential field that drives pedestrians towards it ( static floor field ) ; * the _ obstacles field _ , that indicates for every cell the fact that an obstacle or a wall is within a limited distance , being maximum in cells adjacent to an obstacle and decreasing with the distance ( static floor field ) ; * the _ density field _ that indicates for each cell the pedestrian density in the surroundings at the current time - step , analogous to the concept of _ cumulative mean density _ ( cmd ) indicating the density experienced by pedestrians in a portion of the environment , first introduced by @xcite and elaborated for discrete environments by @xcite ( dynamic floor field ) .
simulation time is modelled in a discrete way by dividing time into steps of equal duration : we assume that a pedestrian moves ( at most , since it is possible to decide to stand still ) 1 cell per time step .
the average velocity of a pedestrian , which can be estimated in real observations or experiments as performed by @xcite in about @xmath5 @xmath6 , will thus determine the duration of the each time step : considering that the size of the cell is @xmath7 , the duration of each time step is @xmath8 .
when running a ca - based pedestrian model , three update strategies are possible @xcite : * _ parallel _ update , in which cells are updated all together ; * _ sequential _ update , in which cells are updated one after the other , always in the same order ; * _ shuffled sequential update _ , in which cells are updated one after the other , but with a different order every time .
the second and third update strategies lead to the definition of asynchronous ca models ( see @xcite for a more thorough discussion on types of asynchronicity in ca models ) . in crowd simulation ca models , parallel update is generally preferred , as mentioned by @xcite , even if this strategy can lead to conflicts that must be solved .
nonetheless , we currently adopted a shuffled sequential update scheme for a first evaluation of the group cohesion mechanism without adding additional mechanisms and parameters to be calibrated for the management of potential conflicts between agents movements .
we focus on two types of group : simple and structured .
simple or informal groups are generally made up of friends or family members and they are characterised by a high cohesion level , moving all together due to shared goals and to a continuous mechanism of adaptation of the chosen paths to try to preserve the possibility to communicate , as discussed by @xcite .
structured groups , instead , are more complex entities , usually larger than simple groups ( more than 4 individuals ) and they can be considered as being composed of sub - groups that can be , in turn , either simple or structured .
structured groups are often artificially defined with the goal of organising and managing a mass movement ( or some kind of other operation ) of a set of pedestrians .
groups can be formally described as : @xmath9 , [ ped_1 , \ldots , ped_n ] \right\rangle\ ] ] structured groups include at least one subgroup , while simple groups only comprise individual pedestrians .
we will refer to the group an agent @xmath10 _ directly _ belongs to as @xmath11 , that is also the smallest group he belongs to ; the _ largest _ group an agent @xmath10 belongs to will instead be referred to as @xmath12 .
it must be noted that @xmath13 only when the agent @xmath10 is member of a simple group that is not included in any structured group .
these sets are relevant for the computation of influences among members of groups of different type ( simple or structured ) . in this model
, a pedestrian is defined as an utility - based agent with state .
functions are defined for utility calculation and action choice , and rules are defined for state - change .
pedestrians are characterised as : @xmath14 where : ( i ) @xmath15 is the agent identification number and @xmath16 is the identification number of the group to which the pedestrian directly belongs to ( for pedestrians that are not member of any group this value is null ) ; ( ii ) @xmath17 is defined as : @xmath18 where @xmath19 indicates the current cell in which the agent is located , and @xmath20 is the direction followed in the last movement ; @xmath21 is the set of possible actions that the agent can perform , essentially movements in one of the eight neighbouring cells ( indicated as cardinal points ) , plus the action of remaining in the same cell ( indicated by an ` @xmath22 ' ) : @xmath23 ; @xmath24 is the goal of the agent in terms of destination area .
this term identifies the current destination of the pedestrian : in particular , every destination is associated to a particular spatial marker .
consequently , @xmath24 is used to identify which path field is relevant for the agent : @xmath25 where @xmath26 is the precise path field associated to @xmath24 and @xmath27 is the path field relevant for the agent .
all these elements take part in the mechanism that manages the movement of pedestrians : agents associate a desirability value , a utility to every movement in the cell neighbourhood , according to a set of factors that concur in the overall decision . in algorithm [ alg1 ]
the agent life - cycle during all the simulation time is proposed : every time step , every pedestrian perceives the values of path field , obstacle field and density field for all the cells that are in its neighbourhood . on the basis of these values and according to different factors , the agent evaluates the different cells around him , associating an utility value to every cell and selects the action for moving into a specific cell .
@xmath28 $ ] @xmath29 @xmath30 @xmath31 @xmath32 \gets evaluation(pf , of , df)$ ] @xmath33)$ ] @xmath34 as previously suggested , the action selection strategy starts gathering the value of floor fields in cells included in the neighbourhood of agent s current position .
the obtained values will be used in the evaluation of the movement towards the related cell . after acquiring the perceived information from the environment
, the agent elaborates a desirability value for each of the admissible actions ( movements ) , according to several factors , that will be described later on .
given the list of possible actions and associated utilities , an action is chosen with a probability proportional to its utility . in particular , the probability for an agent @xmath10 of choosing an action associated to the movement towards a cell @xmath35 is given by the exponential of the utility , normalised on all the possible actions the pedestrian can take in the current turn : @xmath36 where @xmath37 is the normalisation factor and @xmath35 is the currently considered destination cell .
the utility function @xmath38 of a destination cell @xmath35 which corresponds to an action / direction for agent @xmath10 , that takes the form of a weighted sum of components associated to the behavioural specification of an agent : @xmath39 where @xmath40 is the distance of the new cell from the current position , that is 1 for cells in the von neumann neighbourhood ( vertically and horizontally neighbour cells ) and @xmath41 for diagonal cells :
the factor is introduced to penalise the diagonal movements .
the different components of the utility value for a given movement consider the following factors : * the desire to move towards a _ goal _ , a destination in the environment , represented by the @xmath42 function ; * the tendency to stay at a distance from the _ obstacles _
( e.g. walls , columns ) , that are perceived as repulsive , represented by the @xmath43 function ; * the desire to stay at a distance from other individuals , especially those that are not members of the same simple group , an effect of proxemic _ separation _ , represented by the @xmath44 function ; * a _ direction inertia _ factor , increasing the desirability of performing straight forms of movement , represented by the @xmath45 function ; * the penalisation of those movements that cause an _ overlapping _ event , the temporary sharing of the same cell by two distinct pedestrians , represented by the @xmath46 function ; * two contributions related to the tendency to preserve _ group cohesion _ , respectively devoted to simple and structured groups , respectively represented by the @xmath47 and the @xmath48 functions .
note that @xmath49 $ ] : the use of these parameters , in addition to allowing the calibration and the fine tuning of the model , also supports the possibility of describing and managing different types of pedestrian , or even different states of the same pedestrian in different moments of a single simulated scenario . while the above elements are sufficient to generate a pedestrian model that considers the presence of groups , even structured ones , the introduced mechanisms are not sufficient to preserve the cohesion of simple groups , as discussed in a previous work adopting a very similar approach ( see @xcite ) .
this is mainly due to the fact that in certain situations pedestrians adapt their behaviour in a more significant way than what is supported by simple and relatively small modifications of the perceived utility of a certain movement . in certain situations
pedestrians perform an adaptation that appears in a much more decisive way a _ decision _ : they can suddenly seem to temporarily loose interest in what was previously considered a destination to reach and they instead focus on moving closer to ( or at least do not move farther from ) members of their group , generally whenever they perceive that the distance from them has become excessive . in the following , we will discuss a metric of group dispersion that we adopted to quantify this perceived distance and then we will show how it can be used to adapt the weights of the different components of the movement utility computation to preserve group cohesion .
* group dispersion metrics * intuitively , the dispersion of a group can be seen as the degree of spatial distribution of their members . in the area of pedestrian modelling and simulation ,
the estimation of different metrics for group dispersion has been discussed in @xcite in which different approaches are compared to evaluate the dispersion of groups through their movement in the environment . in particular ,
two different approaches are compared here : ( i ) dispersion as occupied area and ( ii ) dispersion as distance from the centroid of the group .
this topic was also considered in the context of computer vision algorithms such as in @xcite , in which however essentially only _ line abreast _ patterns were analysed .
therefore we will focus on the former approach .
formally , the above introduced formulas of group dispersion for each approach are defined as follows : @xmath50 @xmath51 with @xmath52 as the area occupied by the group , @xmath53 as the number of its members , @xmath54 as its centroid .
the second metric appears much more straightforward when a continuous representation of the environment is possible or at least not in contrast with the adopted modelling approach : in the case of a discrete and relatively coarse discretisation its results are not particularly different from the first metric , but they are sometimes counterintuitive especially when describing particular group shapes ( e.g. river - like lanes that are often present in high density situations ) .
the first metric defines the dispersion of the group as the portion of space occupied by the group with respect to the size of the group : the first step of the procedure of computation for this metric builds a convex polygon with the minimum number of edges that contain all the vertices ( representing the position of a pedestrian ) ; the second step computes the area of this polygon .
the dispersion value is calculated as the relationship between the polygon area and the size of the group .
* utility parameters adaptation*[sec : balance ] the adopted approach is characterised by a trade - off process between the goal attraction value and the intra / inter cohesion value in the utility computation : in the situation in which the spatial dispersion value is low , the cohesion behaviour has to influence pedestrian s overall behaviour less than the goal attraction . on the contrary , if the level of dispersion of a group is high , the cohesion component for the members must become more important than the goal attraction .
an adaptation of the two parameters in the utility computation is necessary , by means of a @xmath55 function that can be used to formalise these requirements : @xmath56 where @xmath57 , @xmath58 and @xmath59 are the weighted parameters @xmath38 and @xmath60 is another function that works on the value of group dispersion as the relationship between the area and the size of the group , applying on it the hyperbolic tangent .
the value of @xmath61 is a constant that essentially represents a threshold above which the adaptation mechanism starts to become more influential ; after a face validation phase , we set this value to 2.5 , allowing the output of @xmath62 function in the range @xmath63 $ ] according to all elements in @xmath38 .
the hyperbolic tangent approaches value @xmath64 when @xmath65 approaches @xmath66 ( values @xmath67 indicate a high level of dispersion for small - medium size groups ( 1 - 4 members ) ) . a graphical representation of the trade - off mechanism is shown in fig .
[ fig : balance ] : red and green boxes represent the progress of parameter @xmath59 and parameter @xmath58 ( @xmath57 is treated analogously ) , respectively .
note that the increasing of the dispersion value produces an increment of @xmath59 value and a reduction of @xmath58 parameter .
, for @xmath68 and @xmath69 it must be emphasised the fact that this adaptive balancing mechanism and the current values for its parameters were heuristically established and they actually require a validation ( and plausibly a subsequent calibration ) by comparing results achieved with this configuration and relevant empirical data about group dispersion gathered from actual observations and experiments in controlled situations .
this section describes the results of a simulation campaign carried out to evaluate the performance of the above described model that had mainly two goals : ( i ) _ validate _ the model , in situations for which the adaptation mechanism was not activated ( i.e. no simple groups were present in the simulated population ) , ( ii ) _ evaluate _ the effects of the introduction of simple groups , performing a qualitative face validation of the introduced adaptation mechanism considering available video footages of the behaviour of groups in real and experimental situations .
the chosen situations are relatively simple ones and they were chosen due to the availability of relevant and significant data from the literature .
in particular , we describe here a linear scenario : a _ corridor _ in which we test the capability of the model to correctly reproduce a situation in which two groups of pedestrians enter from one of the ends and move towards the other .
this situation is characterised by a counterflow causing local situations of high density and conflicts on the shared space .
we essentially evaluate and validate this scenario by means of a fundamental diagram as defined by @xcite : it shows how the average velocity of pedestrians in a section ( e.g. one of the ends of a corridor ) varies according to the density of the observed environment . since the flow of pedestrians is directly proportional to their velocity , this diagram is sometimes presented in an equivalent form that shows the variation of flow according to the density . in general
, we expect to have a decrease in the velocity when density grows , due to the growing frequency of `` collision avoidance '' movements by the pedestrians ; the flow , instead , initially grows , since it is also directly proportional to the density , until a certain threshold value is reached ( also called _ critical density _ ) , then it decreases .
despite being of great relevance , different experiments gathered different values of empirical data : while there is a general consensus on the shape of the function , the range of the possible values has even significant differences from different versions .
we investigated this scenario with a significant number of simulations , varying the level of density by adjusting the number of pedestrians present in the environment , so as to analyse different crowding situations .
for every scenario , in terms of environmental configuration and level of density , a minimum of 3 and a maximum of 8 simulations were executed , according to the variability of the achieved results ( more simulations were run when the variability was high , generally around levels of density close to the critical thresholds ) .
every simulation considered at least 1800 simulation turns , corresponding to 10 minutes of simulated time .
the rationale was to observe a good number of complete paths of pedestrians throughout the environment , that was configured to resemble a torus ( e.g. pedestrians completing a movement through it re - entered the scenario from their starting point ) , therefore simulations of situations characterised by a higher density were also set to last longer .
as suggested above , we adopted two different experiment settings : in the first one , the individual pedestrians belonging to a given flow ( i.e. all the pedestrians entering the corridor ) are represented as members of a large structured group , but no simple groups are present .
this first part of the experimentation was also necessary to perform a proper calibration of the model , for the parameters not involved in simple group modelling . in the second experimental setting , we included a variable number of simple groups ( based on the total number of pedestrians in the environment and according to available data on the frequency of groups of different size in a crowd as mentioned in @xcite and @xcite ) first of all to calibrate and qualitatively validate the adequacy of the adaptation mechanism and then to explore its implications on the overall crowd dynamics .
we actually simulated the linear counterflow situation in three different corridors , of growing width : their size is respectively @xmath70 m @xmath71 @xmath72 m ( a ) , @xmath73 m @xmath71 @xmath74 m ( b ) and @xmath75 m @xmath71 @xmath76 m ( c ) .
note that the variation in terms of width and height were applied according to the choice of maintaining the total area at a level about @xmath77 m@xmath78 in every scenario .
a screenshot of the simulation scenario in low density situations including groups of different size is shown in figure [ fig : screenshot ] .
the choice of evaluating the influence of groups in different linear scenarios was also inspired by @xcite , in which a comparison in terms of pedestrians flow from experimental data among three corridors of width @xmath79 m , @xmath70 m and @xmath80 m is presented . in this case , authors show that , in conformance with @xcite , above a certain minimum of about @xmath81 m , the maximum flow is directly proportional to the width of the corridors .
the model was calibrated to achieve , in situations not including simple groups , results in tune with empirical data available from the literature .
we will now focus on some partly counterintuitive results in presence of simple groups . to do so
, data related to the different types of simulated groups were aggregated , and a comparison among the related fundamental diagrams was performed for the movement of groups in corridor a. as summary , fig .
[ fig : av_gruppi ] represents on the same chart all group contributions : the depicted points represent the average flow achieved for that kind of group in the total simulated time , and generally more points are available ( representing different simulation runs ) for the same configuration .
the overall flow of individuals is generally higher than that of groups in almost all situations , and in general with the growth of the size of a simple group we observe a decrease of its overall flow .
moreover , differences tend to decrease and almost disappear after the critical density ( about @xmath82 pedestrians per square metre ) is reached .
we also analysed the effect of the width of the corridor on the flow of groups ( and in general on the overall pedestrian flow ) : figure [ fig : comparison_width ] shows the different fundamental diagrams associated to all simple groups ( irrespectively of their size ) in corridors a , b and c. it is apparent that the critical density moves to higher levels with the growth of the corridor width , in tune with the already mentioned results discussed in @xcite . finally , we analysed how the level of group dispersion , computed by means of the same function employed to manage the adaptive mechanism for group cohesion , varies with a changing density in the environment
. the motivations of this analysis are twofold : first of all , we wanted to understand if the adaptive mechanism for group cohesion is effective , then we wanted to gather empirical data to understand if it produces _ plausible _ results , in line with _ observed data _ that , at the moment of the simulation campaign , were still not available . figure [
fig : dispersion_comparison ] shown the variation of the level of dispersion for groups of size @xmath66 in corridors a , b and c : we can conclude that the mechanism for preserving the cohesion of simple groups is actually effective , since the growth of density does not cause a significant growth of the dispersion . on the other hand , at that moment we could not conclude that the model produces realistic results .
this section comprises several empirical studies aimed at investigating pedestrian crowd dynamics in the natural context by using on - field observation .
in particular the survey was aimed at studying the impact of grouping and proxemics behaviour on the whole crowd pedestrian dynamics .
data analyses were focused on : ( i ) _ level of density and service _ , ( ii ) _ presence of groups _ within the pedestrian flows , ( iii ) group proxemic _ spatial arrangement _ , ( iv ) _ trajectories and walking speed _ of both singles and group members .
furthermore the _ spatial dispersion _ of group members while walking was measured in order to propose an innovative empirical contribution for a detailed description of group proxemics dynamics while walking .
the survey was performed the last 24^th^ of november 2012 from about 2:50 pm to 4:10 pm .
it consisted in the observation of the bidirectional pedestrian flows within the vittorio emanuele ii gallery , a popular commercial - touristic walkway situated in the milan city centre ( italy ) .
the gallery was chosen as a crowded urban scenario , given the large amount of people that pass through it during the weekend for shopping , entertainment and visiting touristic - historical attractions in the centre of milan .
the team performing the observation was composed of four people .
several preliminary inspections were performed to check the topographical features of the walkway .
the balcony of the gallery , that surrounds the inside volume of the architecture from about ten meters in height , was chosen as location thanks to possibility to ( i ) position the equipment for video footages from a quasi - zenithal point of view and ( ii ) to avoid as much as possible to influence the behaviour of observed subjects , thanks to a railing of the balcony partly hiding the observation equipment .
the equipment consisted of two professional full hd video cameras with tripods .
the existing legislation about privacy was consulted and complied in order to exceed ethical issues about the privacy of the people recorded within the pedestrian flows .
two independent coders performed a manual data analyses , in order to reduce errors by crosschecking their results .
a squared portion of the walkway was considered for data analysis : 12.8 meters wide and 12.8 meters long ( 163.84 squared meters ) . in order to perform data analyses ,
the inner space of the selected area was discretised in cells by superimposing a grid on the video ( see fig .
[ fig : grid ] ) ; the grid was composed of 1024 squares 0.4 meters wide and 0.4 meters long .
the bidirectional pedestrian flows ( from north to south and vice versa ) were manually counted minute by minute : 7773 people passed through the selected portion of the vittorio emanuale ii gallery from 2:50 pm to 4:08 pm .
the average level of density within the selected area ( defined as the quantitative relationship between a physical area and the number of people who occupy it ) was detected considering 78 snapshots of video footages , randomly selected with a time interval of one minute .
the observed average level of density was low ( 0.22 people / squared meter ) .
despite it was not possible to analyse continuous situations of high density , several situation of irregular and local distribution of high density were detected within the observed scenario . according to the highway capacity manual by @xcite ,
the level of density in motion situation was more properly estimated taking into account the bidirectional walkway level of service criteria : counting the number of people walking through a certain unit of space ( meter ) in a certain unit of time ( minute ) .
the average level of flow rate within the observed walkway scenario belongs to a _
b _ level ( 7.78 ped / min / m ) that is associated with an irregular flow in low - medium density condition .
the second stage of data analysis was focused on the detection of groups within the pedestrian flows , the number of group members and the group proxemics spatial arrangement while walking .
the identification of groups in the streaming of passerby was assessed on the basis of verbal and nonverbal communication among members : visual contact , body orientation , gesticulation and spatial cohesion among members .
to more thoroughly evaluate all these indicators the coder was actually encouraged to rewind the video and take the necessary time to tell situations of simple local ( in time an space ) similar movements , due to the contextual situation , by different pedestrians from actual group situations .
the whole video was sampled considering one minute every five : a subset of 15 minutes was extracted and 1645 pedestrians were counted ( 21.16% of the total bidirectional flows ) . concerning the flow composition ,
15.81% of the pedestrians arrived alone , while the 84.19% arrived in groups : 43.65% of groups were couples , 17.14% triples and 23.40% larger groups ( composed of four or five members ) .
large structured groups , such as touristic committees , that were present in the observed situation , were analysed considering sub - groups .
results about group proxemics spatial arrangement while walking showed that : * 94.43% of couples was characterised by line - abreast pattern while 5.57% by river - like pattern ; * 31.91% of triples was characterised by line - abreast pattern , 9.57% by river - like pattern and 58.51% by v - like pattern ; * 29.61% of four - person groups is characterised by line - abreast pattern , 3.19% by river - like pattern , 10.39% by v - like pattern , 10.39% triads followed by a single person , 6.23% single individual followed by a triad , 7.79% rhombus - like pattern ( one person heading the group , followed by a dyad and a single person ) , 32.47% of the groups split into two dyads .
the walking speed of both singles and group members was measured considering the path and the time to reach the ending point of their movement in the monitored area ( corresponding to the centre of the cell of the last row of the grid ) from the starting point ( corresponding to the centre cell of the first row of the grid ) .
only the time distribution related to the _ b _ level of service was considered ( as mentioned , the 59% of the whole video footages ) , in order to focus on pedestrian dynamics in situation of irregular flow .
a sample of 122 people was randomly extracted : 30 singles , 15 couples , 10 triples and 8 groups of four members .
the estimated age of pedestrians was approximately between 15 and 70 ; groups with accompanied children were not taken into account for data analyses . about gender , the sample was composed of 63 males ( 56% of the total ) and 59 females ( 44% of the total ) .
differences in age and gender were not considered in this study .
the alphanumeric grid was used to track the trajectories of both single and group members within the walkway and to measure the length of their path ( considering the features of the cells : 0.4 m wide , 0.4 m long ) ( see fig . [
fig : trajectories ] ) .
a first analysis was devoted to the identification of the length of the average walking path of singles ( m=13.96 m , @xmath831.11 ) , couples ( m=13.39 m , @xmath83 0.38 ) , triples ( m=13.34 m , @xmath83 0.27 ) and groups of four members ( m=13.16 m , @xmath83 0.46 ) .
then , the two tailed t - test analyses were used to identify differences in path among pedestrian .
results showed a significant difference in path length between : singles and couples ( p [email protected] ) , singles and triples ( p [email protected] ) , singles and groups of four members ( p [email protected] ) .
no significant differences were detected between path length of couples and triples ( p [email protected] ) , triples and groups of four members ( p [email protected] ) , couples and groups of four members ( p [email protected] ) .
the results showed that the path of singles is 4,48% longer than the average path of group members ( including couples , triples and groups of four members ) .
the walking speed of both singles and group members was detected considering the path of each pedestrian within the flows and the time to reach the ending point from their starting point .
a first analysis was devoted to the identification of the average walking speed of singles ( m=1.22 m / s , @xmath83 1.16 ) , couples ( m=0.92 m / s , @xmath83 0.18 ) , triples ( m=0.73 m / s , @xmath83 0.10 ) and groups of four members ( m=0.65 m / s , @xmath83 0.04 ) .
then , the two tailed t - test analyses were used to identify differences in walking speed among pedestrian .
results showed a significant difference in walking speed between : singles and couples ( p [email protected] ) , singles and triples ( p [email protected] ) , singles and groups of four members ( p [email protected] ) , couples and triples ( p [email protected] ) , triples and groups of four members ( p value@xmath84 0.05 ) . in conclusion
, the results showed that the average walking speed of group members ( including couples , triples and groups of four members ) is 37.21% lower than the walking speed of singles .
the correlated results about pedestrian path and speed showed that in situation of irregular flow singles tend to cross the space with more frequent changes of direction in order to maintain their velocity , avoiding perceived obstacles like slower pedestrians or groups . on the contrary ,
groups tend to have a more stable overall behaviour , adjusting their spatial arrangement and speed to face the contextual conditions of irregular flow : this is probably due to ( i ) the difficulty in coordinating an overall change of direction and ( ii ) the tendency to preserve the possibility of maintaining cohesion and communication among members . in order to improve the understanding of pedestrian proxemics behaviour the last part of the study is focused on the dynamic spatial dispersion of group members while walking .
the dispersion among group members was measured as the summation of the distances between each pedestrian and the centroid ( the geometrical centre of the group ) all normalised by the cardinality of the group .
the centroid was obtained as the arithmetic mean of all spatial positions of the group members , considering the alphanumeric grid . in order to find the spatial positions , the trajectories of the group members belonging to the previous described sample ( 15 couples , 10 triples and 8 groups of four members ) were further analysed . in particular , the positions of the group members were detected analysing the recorded video images every 40 frames ( the time interval between two frames corresponds to about 1.79 seconds , according to the quality and definition of the video images ) starting from the co - presence of the all members on the alphanumeric grid .
this kind of sampling permitted to consider 10 snapshots for each groups .
a first analysis was devoted to the identification of the average proxemics dispersion of couples ( m=0.35 m , @xmath83 0.14 ) , triples ( m=0.53 m , @xmath83 0.17 ) and groups of four members ( m=0.67 m , @xmath83 0.12 ) .
then , the two tailed t - test analyses were used to identify differences in proxemics dispersion among couples , triples and groups of four members .
results showed a significant difference in spatial dispersion between : couples and triples ( p [email protected] ) , couples and groups of four members ( p [email protected] ) .
no significant differences between triples and groups of four members ( p value@xmath85 0.05 ) . in conclusion
, the results showed that the average spatial dispersion of triples and groups of four members while walking is 40.97% higher than the dispersion of couples . in order to be able to provide useful indications for the calibration of the adaptive simple group cohesion mechanism
we also evaluated the average dispersion of the observed groups in terms of area covered by the group : due to the discretisation of the pedestrian localisation mechanism , we were able to essentially count the occupied cells by a sort of convex hull computed considering the positions of pedestrians as vertexes , analogously as for the dispersion metric defined and employed in the simulation model .
the results of this operation estimated the dispersion of couples ( ma=0.6 m@xmath78 ) , triples ( ma=0.8 m@xmath78 ) and groups of four ( ma=1.3 m@xmath78 ) .
these values are currently being employed to calibrate and validate the simple group cohesion mechanism in the conditions of relatively low density and irregular flow .
starting from the achieved results about group proxemics dispersion , we finally focused on a quantitative and detailed description of group spatial layout while walking .
the normalised positions of each pedestrian with respect to the centroid and the movement direction were detected by means of a sample of 10 snapshots for each groups ( 15 couples , 10 triple and 8 groups of four members ) and then further analysed in order to identify the most frequent group proxemics spatial configurations , taking into account the degree of alignment of each pedestrian ( see figure [ fig : scatter - all ] ) .
result showed that couple members tend to walk side by side , aligned to the each other with a distance of 0.4 m ( 36% of the sample ) or 0.8 m ( 24% of the sample ) , forming a line perpendicular to the walking direction ( line abreast pattern ) ; triples tend to walk with a line abreast layout ( 13% of the sample ) , with the members spaced of 0.60 m. regarding groups of four members it was not possible to detect a typical spatial pattern : the reciprocal positions of group members appeared much more dispersed than in the case of smaller groups , probably to due the continuous arrangements in spatial positioning while walking .
the results are in line with the previously described spatial arrangements related to the total observed pedestrian flows ( see section [ sec : flow - comp ] ) , representing an innovative contribution for the understanding of group proxemics dynamics in motion situation , once again in situations characterised by a relatively low density and irregular flow .
empirical data about high density situations would be necessary to actually tune the mechanism in moderate and high density situations , but in this kind of scenario this observation and analysis framework would very likely be inappropriate and ineffective . a more sophisticated and at least partly automated ( employing computer vision techniques and maybe machine learning approaches to support the detection of groups ) controlled experiments ( given the high density situation , some help to the tracking and detection approaches would probably be necessary in terms , for instance , of markers to highlight the heads of group members ) is probably needed to actually face the challenges of acquiring empirical evidences about the behaviour of groups in high density situations .
the participation of psychologists in the definition of such an experimental observation setting would also help in managing any kind of tiring and learning effect due to the experimental procedure .
the paper discussed an integrated approach to the analysis and synthesis of pedestrian and crowd behaviour , in which the two aspects are actually set in an integrated framework and they mutually benefit in different ways . a general schema describing the conceptual pathways connecting modelling and analysis steps was described . a case study in which modelling and simulation approaches , specifically focused on adaptive group behaviour enabling the cohesion of simple groups , produced a research question and some usable techniques to crowd analysis approaches
was also introduced . on the other hand ,
a subsequent observation and analysis about the phenomenology represented in the model was also described .
currently the empirical evidences resulting from this analysis activity are being used to validate and calibrate the model for group cohesion . in particular the simulation model
correctly reproduces some of the observed phenomena , in particular , lower walking speeds for groups and tendency to preserve cohesion ( although this aspect is undergoing a further calibration ) .
additional elements that are now being evaluated are related to the capability of the model to generate spatial patterns resulting from the analysis . moreover , the analysis of the gathered video footage also highlighted additional phenomenologies that are now being more thoroughly analysed : in particular , patterns of `` leader follower '' behaviour within groups were detected and the introduced simulation model presents all the necessary elements and mechanisms to represent this kind of pattern .
future works are also aimed at supporting innovative automated analysis techniques based on computer vision and machine learning approaches .
the authors would like to thank milano s municipality for granting the necessary authorisations to carry out the observation in the vittorio emanuele ii gallery .
bandini , s. , rubagotti , f. , vizzari , g. , shimura , k. , 2011 .
an agent model of pedestrian and group dynamics : experiments on group cohesion . in : pirrone , r. , sorbello , f. ( eds . ) ,
ai*ia . vol .
6934 of lecture notes in computer science .
springer , pp .
104116 .
federici , m. l. , gorrini , a. , manenti , l. , vizzari , g. , 2012 .
data collection for modeling and simulation : case study at the university of milan - bicocca . in : sirakoulis ,
g. c. , bandini , s. ( eds . ) , acri .
7495 of lecture notes in computer science .
springer , pp .
699708 .
georgoudas , i. g. , sirakoulis , g. c. , andreadis , i. , 2011 .
an anticipative crowd management system preventing clogging in exits during pedestrian evacuation processes .
ieee systems journal 5 ( 1 ) , 129141 .
leal - taix , l. , pons - moll , g. , rosenhahn , b. , 2011 .
everybody needs somebody : modeling social and grouping behavior on a linear programming multiple people tracker . in : iccv workshops .
ieee , pp .
120127 .
manzoni , s. , vizzari , g. , ohtsuka , k. , shimura , k. , 2011 . towards an agent - based proxemic model for pedestrian and group dynamics : motivations and first experiments . in : tumer ,
yolum , sonenberg , stone ( eds . ) , proc . of 10th int .
conf . on autonomous agents and multiagent systems innovative applications track ( aamas 2011 ) .
12231224 .
milazzo ii , j. s. , rouphail , n. m. , hummer , j. e. , allen , d. p. , 1999 .
quality of service for interrupted - flow pedestrian facilities in highway capacity manual 2000 .
transportation research record : journal of the transportation research board 1678 ( -1 ) , 2531 .
moussad , m. , perozo , n. , garnier , s. , helbing , d. , theraulaz , g. , 04 2010 .
the walking behaviour of pedestrian social groups and its impact on crowd dynamics .
plos one 5 ( 4 ) , e10047 .
http://dx.doi.org/10.1371%2fjournal.pone.0010047 raghavendra , r. , bue , a. d. , cristani , m. , murino , v. , 2011 . abnormal crowd behavior detection by social force optimization . in : salah , a. a. , lepri , b. ( eds . ) , hbu .
7065 of lecture notes in computer science .
springer , pp .
134145 .
schadschneider , a. , klingsch , w. , klpfel , h. , kretz , t. , rogsch , c. , seyfried , a. , 2009 .
evacuation dynamics : empirical results , modeling and applications . in : meyers , r. a. ( ed . ) , encyclopedia of complexity and systems science .
springer , pp .
31423176 .
schultz , m. , rger , l. , fricke , h. , schlag , b. , 2012 .
group dynamic behavior and psychometric profiles as substantial driver for pedestrian dynamics . in : pedestrian and evacuation dynamics
2012 ( ped2012 ) .
http://arxiv.org/abs/1210.5553 schultz , m. , rger , l. , fricke , h. , schlag , b. , 2012 .
group dynamic behavior and psychometric profiles as substantial driver for pedestrian dynamics . in : pedestrian and evacuation conference ( ped2012 ) .
http://arxiv.org/abs/1210.5553 vizzari , g. , manenti , l. , ohtsuka , k. , shimura , k. , 2012 .
an agent - based approach to pedestrian and group dynamics : experimental and real world scenarios . in : proceedings of the 7th international workshop on agents in traffic and transportation .
http://www.ia.urjc.es/att2012/papers/att2012_submission_1.pdf was , j. , 2010 .
crowd dynamics modeling in the light of proxemic theories . in : rutkowski , l. , scherer , r. , tadeusiewicz , r. , zadeh , l. a. , zurada , j. m. ( eds . ) , icaisc ( 2 ) . vol .
6114 of lecture notes in computer science .
springer , pp . 683688 .
weidmann , u. , 1993 .
transporttechnik der fussgnger - transporttechnische eigenschaftendes fussgngerverkehrs ( literaturstudie ) .
literature research 90 , institut fer verkehrsplanung , transporttechnik , strassen- und eisenbahnbau ivt an der eth zrich .
zhang , j. , klingsch , w. , schadschneider , a. , seyfried , a. , 2011 .
transitions in pedestrian fundamental diagrams of straight corridors and t - junctions .
journal of statistical mechanics : theory and experiment 2011 ( 06 ) , p06004 . | studies related to _ crowds _ of pedestrians , both those of theoretical nature and application oriented ones , have generally focused on either the _ analysis _ or the _ synthesis _ of the phenomena related to the interplay between individual pedestrians , each characterised by goals , preferences and potentially relevant relationships with others , and the environment in which they are situated . the cases in which these activities have been systematically integrated for a mutual benefit are still very few compared to the corpus of crowd related literature .
this paper presents a case study of an integrated approach to the definition of an innovative model for pedestrian and crowd simulation ( on the side of synthesis ) that was actually motivated and supported by the analyses of empirical data acquired from both experimental settings and observations in real world scenarios .
in particular , we will introduce a model for the adaptive behaviour of pedestrians that are also members of groups , that strive to maintain their cohesion even in difficult ( e.g. high density ) situations .
the paper will show how the synthesis phase also provided inputs to the analysis of empirical data , in a virtuous circle .
crowd analysis , crowd synthesis , agent based modelling and simulation , groups |
in recent years , cosmology has converged on a `` concordance '' model that continues to meet with remarkable success wherever new observational tests are introduced .
the direct data span a large dynamic range of structural scales , including power spectrum probes over about a factor of @xmath0 in linear comoving scale .
the microwave background spectrum and the products of cosmological nucleosynthesis confirm basic model assumptions and microphysics back to the first fraction of a second of the big bang .
an important part of the standard scenario is inflation .
the generic inflationary predictions of near global flatness and nearly scale - invariant , gaussian fluctuations are well confirmed by the data .
we thus seem to have some direct information , on very large comoving scales , dating from some very early time , prior to the terascale ( or tev energies ) , when the hubble length was less than about a millimeter and the hubble time less than a few picoseconds . at the same time
, advances in string theory have opened up a wide range of new possibilities for the initial conditions and the behavior of the universe at early times .
seven extra dimensions seem mandatory for consistency ; they must be compactified , or gravity must be localized by some other means ; depending on the order of events , these possibilities may require extensions to the usual framework of inflation based on field theory in 3 + 1 dimensions .
initial conditions , and even apparently fundamental parameters such as the shapes of effective potentials in the field theory lagrangian , may be strongly shaped by anthropic selection from a string - theory `` landscape '' , changing the character of arguments about what is natural for inflationary scenarios .
branes of various dimensionalities introduce new behavior not present in a system of fields in 3d space , including new possibilities for inflation itself , for the generation of fluctuations , and for new kinds of macroscopic defects such as cosmic superstrings .
such a large range of new possibilities can be introduced because the observational tests of concordance cosmology can not directly probe a very large swathe of early cosmic history .
viewed in terms of logarithmic scales in energy space or cosmic scale factor , the unexplored region of our past light cone greatly exceeds the explored region .
for example , information from particle content only survives from the decoupling of the relevant species ; for most dark matter candidates , this is well below a gev in temperature .
the baryon content is a presumed relic of the pre - tev period , which sets the only firm lower limit on the end of inflation . and
while it is likely true that the largest - scale perturbations date from very early , many ( possibly as many as about 27 ) orders of magnitude of expansion before the end of inflation , the direct data on these cover only a relatively brief period and a few orders of magnitude of comoving scale .
we have almost no direct data on what happened during about 30 orders of magnitude of cosmic expansion ; for almost all of this period , we do not even have solid theoretical arguments to believe that we understand the dimensionality of space .
advances in detection of high frequency gravitational waves are starting to provide a way to explore observationally this very large range of scales , epochs and phenomena not previously accessible . a new generation of ground - based detectors operating in the sub - kilohertz band@xcite , such as ligo , geo , and virgo ,
are now collecting data at their design sensitivity , and will be significantly upgraded in the years ahead .
the space interferometer@xcite lisa , operating in the millihertz band , will have its core technology proven in a test flight in 2009 and will itself fly in about ten years .
pulsar surveys are providing larger samples of clean pulsars for timing - based detection at frequencies of inverse years to decades@xcite .
together , these techniques are maturing to the point where we will have cosmologically sensitive information over a range of frequencies from kilohertz to nanohertz . since gravitational waves
penetrate all of cosmic history at sub - planckian densities , and even propagate some distance into extra dimensions , these techniques will access directly many phenomena that are currently hidden from view .
this brief survey represents an introduction and update ; more detail on many of the ideas presented here appear in earlier reviews and surveys@xcite .
also , since the emphasis is on classically - generated waves that might be directly detected in the near future ( and not on cosmic background techniques ) , inflationary quantum perturbations are little discussed , even though they dominate calculations in the cosmological literature ; direct detection of these waves might occur in the more distant future@xcite .
an informative plot of all cosmic history is a `` redshifted hubble frequency diagram '' , shown in figure 1 .
the inverse apparent horizon size at a scale factor @xmath1 , or hubble frequency @xmath2 , is redshifted today to a lower frequency , @xmath3 .
we plot this observed frequency as a function of @xmath1 ( in units where the current scale factor @xmath4 ) .
several simple scalings are : @xmath5 or slow roll inflation corresponds to @xmath6 ; matter - dominated evolution gives @xmath7 ; radiation - dominated evolution gives @xmath8 .
a standard inflationary history with inflation at the gut scale is shown as the uppermost solid line .
inflation ending at much lower energy is shown as other solid lines .
the upper dashed lines show a reference universe that stays permanently radiation - dominated.@xcite for reference , note that a frequency of about 0.1 millihertz today corresponds to the terascale a temperature of about 1tev , and an apparent horizon size of about 1 mm : @xmath9 [ fig : figure1 ] the cross - hatched regions show areas about which we currently have at least some fairly direct data . at the lower right
, we have the universe since nucleosynthesis ; at the lower left , standard gut scale inflation mapping onto the cmb anisotropies . the diagram has the virtue of expanding much intermediate history over a wide range of scales . the figure can be regarded as applying to extra dimensions as well , very early when these may be in play , as long as they can be described by a single scale factor .
horizontal lines in this plot show constant observed frequency ; the universe lines cross these at observed hubble frequencies in a given model ; above this corresponds to frequencies within the horizon ( that is , less than @xmath10 ) at a given @xmath1 .
lines are shown corresponding to frequency bands for ground - based interferometers , space - based interferometers , and pulsar timing techniques .
the shaded regions correspond to areas where directly observable cosmic gravitational wave backgrounds might be generated .
those to the right of the peak correspond to classical processes ; those to the left correspond to quantum processes .
all the current probes revealing the character of relic structure via cmb , large scale structure , and quasar absorption extend only four or five orders of magnitude up from the bottom of the graph . constraints on cosmic history from the cmb spectrum and nucleosynthesis extend farther , back to weak decoupling .
the formation of wimp - type dark matter and neutrino decoupling happen at about the same time , and axion dark matter condenses at the qcd epoch . from earlier times , microscopic information tends to be thermalized , apart from a few conserved quantities ; indeed , perhaps even baryon number does not survive from far above the terascale .
events in the top two - thirds of the figure , including most of cosmic expansion history ( in log units ) , are currently invisible .
that allows for a very wide range of possible new physics .
gravitational waves have the potential to open up much of this unexplored region to direct observation .
they will at least constrain models in a meaningful way , and may reveal new sources of gravitational waves .
there is a possibility of obtaining detailed information about cosmic activity on mesoscopic scales , and about what physics was up to during those `` invisible ages '' .
[ fig : figure2 ] = 3.5 in
a schematic spectrum of the gravitational wave background is shown in figure 2 .
an optimistic estimate of conventional inflationary waves is shown , saturating the current upper limit on tensor / scalar ratio , little damping , and a relativistic equation of state .
the current limit on the tensor / scalar ratio limits these to about @xmath11 of the cmb energy density , but in many inflation models it is much less@xcite ; roughly speaking , the metric perturbation on the horizon is of the order of the inflationary expansion rate in planck units .
it is not that surprising that the backgrounds are weak , since the production process corresponds to about a single graviton quantum per hubble 4-volume during inflation ; the occupation numbers are still at best of the order of unity when they arrive at our detectors .
( however , note that there are less conventional versions of inflation , particularly pre - big - bang scenarios , with quite different and sometimes quite intense predicted spectra at high frequencies , again illustrating how much there is to learn from the gravitational wave backgrounds ; see for example@xcite ) .
the new classical sources discussed here are potentially much stronger in certain ranges of frequency ; the natural scale of energy density for these is the cmb density , times inefficiency factors which are smaller than unity by a factor depending on powers of the scale relative to the hubble length .
note that lisa can detect backgrounds down to about six orders of magnitude less energy than the relativistic plasma , so it can detect sources with net gravitational radiation efficiency as small as about @xmath12 .
more explanation of various parts of this figure can be found in reviews and summaries cited above ; a more detailed discussion for lisa in particular , exploiting the use of sagnac calibration for extracting broad band backgrounds , is in@xcite . during an early first order phase transition ,
the universe `` boils'' it nucleates bubble of new phase , and the growth of bubbles converts internal energy ( the latent heat of transition ) to relativistic flows on the nucleation scale . for a strong transition
the separation between nucleated bubbles is macroscopically large , leading to flows that are coherent on scales about one to two orders of magnitude smaller than the horizon size , a dimensionless ratio which determines the characteristic peak frequency and radiation efficiency .
the flows generate modestly relativistic bulk turbulent velocities and accelerate fluid mass , leading to generation of gravitational waves .
the kinetic energy dissipates to smaller scales in a turbulent cascade , creating a power law of higher frequency radiation .
gravitational wave spectra were estimated for qcd@xcite and electroweak@xcite phase transitions during the 1980 s .
these papers estimated basic parameters such the characteristic frequency and intensity of the backgrounds , which depend mainly on the critical temperature and the degree of supercooling ( determined by the latent heat of transition and the nucleation process , e.g. @xcite ) .
further work during the 1990 s down to the present @xcite made more detailed inroads into accurate models of the spectra , although these are still not realistic or definitive , as one might expect given the difficulties in understanding turbulent processes .
figure 3 shows an enlarged view of the lisa band , with a typical peak frequency and amplitude shown .
note the importance of understanding the astrophysical foregrounds in this frequency range with similar confusion - limited spectra , especially galactic and extragalactic dwarf binaries@xcite ; this will likely be the key limiting factor of the sensitivity to high redshift transitions @xcite .
progress has continued on theoretical understanding of the physics of the transition itself.@xcite recent interest is especially prompted by the approach of real data on the terascale soon to come from cern s large hadron collider .
lisa s band is well tuned to phase transitions at lhc scales since its peak sensitivity corresponds to about a tenth of the horizon scale at 1 tev .
it has been proposed that the elecroweak transition is responsible for the departure from equilibrium that brought about baryogenesis ; these scenarios could be directly confirmed by a lisa detection
. string theory requires many extra dimensions .
the sizes of the dimensions , their shapes , and how they are stabilized are unknown .
all we really know is that for string theory to be right , the extra dimensions must be small , or their radii of curvature must be small , so that they do not appear today in direct tests of the gravitational inverse - square law at submillimeter scale@xcite .
( the scales probed by standard model fields are of course much smaller than this , but they might be confined to a 3-brane living in a larger dimensional space . ) since it turns out the hubble length at the terascale is also about a millimeter , the current threshold of ignorance happens to be about the same in the laboratory gravity , particle / field , and cosmological realms .
thus we find lab experiments , accelerator physics , and lisa cosmology converging on the same new regime in very different ways .
it is even possible that new properties of gravity on this scale are related to cosmic dark energy ( whose energy density is about @xmath13 ) .
the formation of extra stable dimensions introduces sources of free internal energy that might be released coherently on a macroscopic scale.@xcite moreover it introduces another kind of mechanism for generating gravitational waves : motion and curvature of our standard model brane in the extra dimensions . at the time of the localization of the graviton by curving extra dimensions ,
the position of our brane might be a random variable above the horizon / curvature scale , and in general , spatially inhomogeneous modes would be excited by the kibble mechanism .
( note that most braneworld scenarios concentrate instead on generation of quantum noise in an already - stabilized brane system ; see e.g. @xcite ) alternatively is also possible that the new scales are all much smaller than the hubble length when the stabilization occurs . in this case
, the behavior of the extra dimensions can be described by a scalar order parameter as a function of 3 + 1d position , and the effect in 3 + 1d spacetime is similar to the phase transitions just discussed .
there is no mandatory reason to assign an extremely high redshift to inflation .
reheating ends as late as the terascale in many scenarios ( particularly those braneworlds where the planck scale is not far above the terascale ) , and could have ended as late as the qcd epoch or even electroweak decoupling without directly affecting current cosmological observables .
if we are fortunate , it might have ended with enough classical gravitational noise to be accessible to direct gravitational wave detection .
there is also no fundamental reason to assume a very quiet reheating .
inflation itself is an extraordinarily coherent behavior of a scalar field ; reheating is a process that eventually converts its internal potential energy into a thermal mix of relativistic particles . in many scenarios ( especially `` hybrid '' ones )
, the conversion begins with macroscopically coherent but inhomogeneous motions that eventually cascade to microscopic scales .
quantum coherent processes such as `` preheating '' transform into coherent classical motions which , like the phase transitions discussed above , generate backgrounds of the order of @xmath14 of the thermal plasma density@xcite .
( as with those transitions , the frequency of the background only matches the gravitational wave detectors if the final activity of this type occurs well below the gut scale ) .
a closely related effect is the formation of nonlinear horizon - scale domains in axion- or higgs- like fields of very low mass ; in this case the field s internal energy is converted not into radiation , but into miniclusters of cosmic dark matter@xcite .
the spectra from the above three sources resemble each other : a broad peak around a characteristic frequency , a power law tail at high frequency from the cascade , and a steep rolloff at low frequency from causality ( the redshifted hubble frequency ) , since low frequencies only come from small velocity flows which are inefficient radiators of gravitational waves .
the above examples show that lisa is positioned to detect direct evidence of first - order phase transitions , or indeed any significant sharing of internal energy in sub - horizon - scale bulk motions near and somewhat above the terascale .
cosmic strings have been studied for many years as a possible new form of mass - energy with new and distinctive astrophysical effects@xcite they were originally conceived in field theory as defects caused by broken u(1 ) symmetries in yang - mills theories .
recently they have re - emerged as possible quasi - stable structures of fundamental string theory , sometimes called cosmic superstrings @xcite that tend to arise naturally as u(1 ) symmetries are broken in models of brane inflation @xcite . a tangled net of superstrings forms by a process of kibble quenching after cosmological inflation .
the primordial net of long strings continually intersects with itself , spawning isolated , oscillating loops that ultimately radiate almost all of their energy into gravitational waves@xcite .
although the fundamental physics differs widely for different types of strings , their quantitative gravitational effects are mainly governed by one fundamental parameter , the dimensionless mass per length or tension @xmath15 ( in planck units with @xmath16 ) .
current limits on gravitational wave backgrounds ( from pulsar timing ) already suggest that if superstrings exist , they must be so light that they have no observable astrophysical effects other than their gravitational radiation .
figure 4 shows predicted stochastic background spectra@xcite from strings for various values of @xmath15 .
the current pulsar limits corresponding to @xmath17 already significantly constrain brane cosmology , and lisa will probe beyond this limit by several orders of magnitude , to @xmath18 . the spectrum from superstrings
is clearly distinguishably different from that of phase transitions or any other predicted source : it is nearly flat ( in @xmath19 units ) over many decades at high frequencies , including the range where lisa is likely to observe it .
there is a possibility , if the strings are not too much lighter than current limits , that occasional distinctive bursts might be seen from nearby loops that happen to beam gravitational waves in our direction from cusp catastrophes in the loop trajectory@xcite .
these rare events , if they are intense enough to stand out above the background , are recognizable in principle from their universal waveform , which derives just from the geometry of the cusps . approaching time @xmath20 from the moment of the catastrophe , in units given by the fundamental mode of the loop , the metric strain amplitude due to radiation beamed from a cusp varies like @xmath21 , and
is beamed within an angle @xmath22 .
( that is , if observed at angle @xmath23 from the beam direction , the cusp waveform behavior is smoothed out at @xmath24 . )
although individual burst events , if detected , give the clearest signature of a string source , the first detectable sign of a superstring loop population is likely their integrated confusion - limited stochastic background@xcite . while lisa s upper limits will certainly provide interesting constraints and eliminate classes of cosmological theories , the actual discovery of an identifiable superstring background ( and even better , possible but less likely , superstring bursts ) would be a direct observation of a stringy signature in nature .
measured properties would provide insight into physics beyond field theory and classical general relativity , and into the physics underlying cosmic inflation .
real data would go a long way to help shape the rich mathematical insights of string theory into a model of the real world .
v. m. kaspi , j. h. taylor and m. f. ryba , astrophys .
j. * 428 * , 713 ( 1994 ) . s. e. thorsett and r. j. dewey , phys .
d * 53 * , 3468 ( 1996 ) .
m. p. mchugh , g. zalamansky , f. vernotte and e. lantz , phys .
d * 54 * , 5993 ( 1996 ) .
a. n. lommen and d. c. backer , astrophys . j. * 562 * , 297 ( 2001 ) [ arxiv : astro - ph/0107470 ] .
a. n. lommen , `` new limits on gravitational radiation using pulsars , '' arxiv : astro - ph/0208572 . c. j. hogan , `` cosmological gravitational wave backgrounds , '' arxiv : astro - ph/9809364 . m. maggiore , phys .
rept . * 331 * , 283 ( 2000 ) [ arxiv : gr - qc/9909001 ] . c. cutler and k. s. thorne , `` an overview of gravitational - wave sources , '' arxiv : gr - qc/0204090 .
s. a. hughes , annals phys . * 303 * , 142 ( 2003 ) [ arxiv : astro - ph/0210481 ] .
a. buonanno , `` gravitational waves from the early universe , '' arxiv : gr - qc/0303085 .
s. chongchitnan and g. efstathiou , phys .
d * 73 * , 083511 ( 2006 ) [ arxiv : astro - ph/0602594 ] .
g. efstathiou and s. chongchitnan , `` the search for primordial tensor modes , '' arxiv : astro - ph/0603118 .
l. a. boyle and p. j. steinhardt , `` probing the early universe with inflationary gravitational waves , '' arxiv : astro - ph/0512014 .
t. l. smith , m. kamionkowski and a. cooray , `` direct detection of the inflationary gravitational wave background , '' arxiv : astro - ph/0506422 .
j. crowder and n. j. cornish , phys .
d * 72 * , 083005 ( 2005 ) [ arxiv : gr - qc/0506015 ] .
j. d. bjorken , `` the classification of universes , '' arxiv : astro - ph/0404233 .
a. buonanno , m. maggiore and c. ungarelli , phys .
d * 55 * , 3330 ( 1997 ) [ arxiv : gr - qc/9605072 ] .
c. j. hogan and p. l. bender , phys .
d * 64 * , 062002 ( 2001 ) [ arxiv : astro - ph/0104266 ] .
a. kosowsky , m. s. turner and r. watkins , phys .
* 69 * , 2026 ( 1992 ) .
a. kosowsky , m. s. turner and r. watkins , phys .
d * 45 * , 4514 ( 1992 ) .
a. kosowsky and m. s. turner , phys .
d * 47 * , 4372 ( 1993 ) [ arxiv : astro - ph/9211004 ] .
m. kamionkowski , a. kosowsky and m. s. turner , phys .
d * 49 * , 2837 ( 1994 ) [ arxiv : astro - ph/9310044 ] .
a. kosowsky , a. mack and t. kahniashvili , phys .
d * 66 * , 024030 ( 2002 ) [ arxiv : astro - ph/0111483 ] .
a. d. dolgov , d. grasso and a. nicolis , phys .
d * 66 * , 103505 ( 2002 ) [ arxiv : astro - ph/0206461 ] .
r. apreda , m. maggiore , a. nicolis and a. riotto , nucl .
b * 631 * , 342 ( 2002 ) [ arxiv : gr - qc/0107033 ] .
a. nicolis , class .
* 21 * , l27 ( 2004 ) [ arxiv : gr - qc/0303084 ] . c. grojean and g. servant ,
gravitational waves from phase transitions at the electroweak scale and arxiv : hep - ph/0607107 . c. caprini and r. durrer , `` gravitational waves from relativistic sources : primordial turbulence and magnetic fields , '' arxiv : astro - ph/0603476 .
a. j. farmer and e. s. phinney , mon . not .
soc . * 346 * , 1197 ( 2003 ) [ arxiv : astro - ph/0304393 ] .
d. coward and t. regimbau , `` detection regimes of the cosmological gravitational wave background from astrophysical sources , '' arxiv : astro - ph/0607043 .
k. kajantie , m. laine , k. rummukainen and m. e. shaposhnikov , phys . rev . lett . *
77 * , 2887 ( 1996 ) ; k. rummukainen , m. tsypin , k. kajantie , m. laine and m. e. shaposhnikov , nucl .
b * 532 * , 283 ( 1998 ) ; f. csikor , z. fodor and j. heitger , phys .
lett . * 82 * , 21 ( 1999 ) . c. grojean , g. servant and j. d. wells , phys .
d * 71 * , 036001 ( 2005 ) [ arxiv : hep - ph/0407019 ] . c. d. hoyle , d. j. kapner , b. r. heckel , e. g. adelberger , j. h. gundlach , u. schmidt and h. e. swanson , phys .
d * 70 * , 042004 ( 2004 ) [ arxiv : hep - ph/0405262 ] .
c. j. hogan , phys .
lett . * 85 * , 2044 ( 2000 ) [ arxiv : astro - ph/0005044 ] .
c. j. hogan , phys .
d * 62 * , 121302 ( 2000 ) [ arxiv : astro - ph/0009136 ] .
n. jones , h. stoica , and s. h. h. tye , jhep * 07 * , 051 ( 2002 ) , hep - th/0203163 . s. sarangi and s. h. h. tye , phys .
b536 * , 185 ( 2002 ) , hep - th/0204074 .
n. t. jones , h. stoica , and s. h. h. tye , phys .
b563 * , 6 ( 2003 ) , hep - th/0303269 .
s. kachru , r. kallosh , a. linde , j. m. maldacena , l. mcallister and s. p. trivedi , jcap * 0310 * , 013 ( 2003 ) [ arxiv : hep - th/0308055 ] .
r. durrer , aip conf .
proc . * 782 * , 202 ( 2005 ) [ arxiv : hep - th/0507006 ] .
h. firouzjahi and s. h. tye , jcap * 0503 * , 009 ( 2005 ) [ arxiv : hep - th/0501099 ] .
t. hiramatsu , phys .
d * 73 * , 084008 ( 2006 ) [ arxiv : hep - th/0601105 ] .
l. kofman and p. yi , phys .
d * 72 * , 106001 ( 2005 ) [ arxiv : hep - th/0507257 ] .
g. n. felder and l. kofman , `` nonlinear inflaton fragmentation after preheating , '' arxiv : hep - ph/0606256 .
r. easther and e. a. lim , jcap * 0604 * , 010 ( 2006 ) [ arxiv : astro - ph/0601617 ] . k. m. zurek , c. j. hogan and t. r. quinn , `` astrophysical effects of scalar dark matter miniclusters , '' arxiv : astro - ph/0607341 .
t. w. b. kibble , j. phys . a 9 , 1387 ( 1976 ) y. b. zeldovich ,
mon . not .
. soc . * 192 * , 663 ( 1980 ) .
a. vilenkin , phys .
lett . * 46 * , 1169 ( 1981 ) .
a. vilenkin , phys . lett . *
107b * , 47 ( 1981 ) m. b. hindmarsh and t. w. b. kibble , rept .
. phys . * 58 * , 477 ( 1995 ) , hep - ph/9411342 .
a. vilenkin , e. p. s. shellard , _ cosmic strings and other topological defects _ , cambridge university press ( 2000 )
e. j. copeland , r. c. myers and j. polchinski , jhep * 0406 * , 013 ( 2004 ) [ arxiv : hep - th/0312067 ] .
m. g. jackson , n. t. jones and j. polchinski , jhep * 0510 * , 013 ( 2005 ) [ arxiv : hep - th/0405229 ] .
j. polchinski , `` introduction to cosmic f- and d - strings , '' arxiv : hep - th/0412244 ( 2004 ) j. polchinski , aip conf .
proc . * 743 * , 331 ( 2005 ) [ int .
j. mod .
a * 20 * , 3413 ( 2005 ) ] [ arxiv : hep - th/0410082 ] .
a. c. davis , t. w. b. kibble , contemp .
46 , 313 ( 2005 ) t. w. b. kibble , `` cosmic strings reborn ?
, '' arxiv : astro - ph/0410073 .
a. vilenkin , `` cosmic strings : progress and problems , '' arxiv : hep - th/0508135 .
c. j. hogan , phys .
d * 74 * , 043526 ( 2006 ) [ arxiv : astro - ph/0605567 ] .
t. damour and a. vilenkin , phys .
d * 71 * , 063510 ( 2005 ) [ arxiv : hep - th/0410222 ] .
x. siemens , j. creighton , i. maor , s. ray majumder , k. cannon and j. read , phys .
d * 73 * , 105001 ( 2006 ) [ arxiv : gr - qc/0603115 ] . | forthcoming advances in direct gravitational wave detection from kilohertz to nanohertz frequencies have unique capabilities to detect signatures from or set meaningful constraints on a wide range of new cosmological phenomena and new fundamental physics .
a brief survey is presented of the post - inflationary gravitational radiation backgrounds predicted in cosmologies that include intense new classical sources such as first - order phase transitions , late - ending inflation , and dynamically active mesoscopic extra dimensions .
lisa will provide the most sensitive direct probes of such phenomena near tev energies or terascale .
lisa will also deeply probe the broadband background , and possibly bursts , from loops of cosmic superstrings predicted to form in current models of brane inflation .
address= astronomy and physics departments , university of washington , seattle , washington 98195 - 1580 |
the principal aim of this paper is to extract information on the fluxes of be and cno neutrinos directly from solar neutrino experiments , with minimal assumptions about solar models . in this respect
, we will update previous results @xcite and try to elucidate the role of cno neutrinos .
we will see that experimental data are more and more against the hypothesis of standard neutrinos ( without mass , mixing , magnetic moments ... ) .
next we will compare these informations with solar models , both standard and non standard ones .
clearly , low ( smaller than standard ) central temperature models are ruled out , essentially because they can not reproduce the experimental data available on both be and b neutrinos .
hybrid models , where some suitable nuclear cross section is varied in order to reduce the be neutrinos flux to the observed value and with a higher central temperature , so as to agree with experimental results on b neutrinos flux , can also be excluded , as in these models the cno neutrino flux grows beyond acceptable levels . in other words ,
the bounds on be ( cno ) neutrinos tell us that it is hopeless to reduce ( enhance ) the central solar temperature , in order to stay with standard neutrinos . in summary
, we shall demonstrate that , under the assumption of standard neutrinos : * the available experimental results look inconsistent among themselves , even if one of the four experiments were wrong ; * the flux of intermediate energy neutrinos ( be+cno ) as derived from experiments is significantly smaller than the prediction of ssm s ; * the different reduction factors for @xmath0be and @xmath1b neutrinos with respect to the ssm are essentially in contradiction with the fact that both @xmath0be and @xmath1b neutrinos originate from the same parent @xmath0be nucleus .
we will discuss then the expectations for borexino , both in the case of standard and non standard neutrinos , showing that the experiment can clearly discriminate among several possible solutions to the solar neutrino puzzle . ? = ?
[ cols= " < , > , < , > , < , < , > , < , < " , ]
one could imagine the conspiracy of two mechanisms , so as to bring both @xmath2 and @xmath3 in agreement with experiment .
for example , one could assume that @xmath4 is much larger than commonly assumed ( e.g. as a result of a hypothetical resonance @xcite ) so as to enhance the ppi channel and reduce @xmath2 to the desired value .
at the same time , by varying some suitable parameter the central temperature could be increased , so as to bring @xmath3 in agreement with experiment .
this mechanism also fails @xcite , see fig .
[ fig3 ] , due to the fact that as temperature raises , the cno flux grows as fast as the boron flux , and the experimental bound on @xmath5 is again violated .
in other words , while beryllium and boron neutrinos tell us that one can not hope to solve the neutrino problem by lowering the central temperature , the bound on cno implies that increasing the temperature does not work either .
if instead the temperature is unchanged but @xmath6 is increased , one has still the problem that the ssm prediction for the cno exceeds the experimental constraint , see again fig .
[ fig3 ] and ref .
we have seen that , for standard neutrinos , the be - flux is strongly suppressed with respect to the ssm predictions . what has to be expected for non standard neutrinos ?
in table [ table2 ] we update and extend a recent analysis @xcite for a few candidate solutions .
we use now as a reference the fluxes corresponding to the `` best model with helium and metal diffusion '' of ref . @xcite . for active neutrinos ,
both small and large angle msw solutions are acceptable , as well as the just - so model . on the other hand , for sterile neutrinos only msw at small angle
gives a good fit . among these four acceptable solutions , two of them ( msw large angle and just - so )
give signals ( cc+nc ) that are quite a significant fraction of the ssm prediction , see last column in table [ table2 ] .
in other words , in face of the present experimental data , the beryllium signal does not need to be small , for non standard neutrinos .
the situation is made more clear in fig .
[ fig4 ] , where we show the 90% c.l .
regions according to the different models .
a direct measurement of the be line can in many cases discriminate among the possibile solutions .
very large signals , above 75% of the ssm prediction , correspond essentially to the just - so solution .
between 75% and about 35% various models are acceptable .
between 35% and 20% the solution has to bee msw at small angle for active neutrinos .
very small signal , say below 20% , are only possible for standard neutrinos , or transitions into sterile neutrinos .
most of the results presented here are the outcome of the fruitful and friendly collaboration in the last few years with v. berezinsky , v. castellani , f. ciacio , s. deglinnocenti and n. ferrari .
we are grateful to t. e. o. ericson and to g. cocconi for useful discussions .
one of us ( g. f. ) thanks the cern theory division for hospitality while this work was done . | we extract information on the fluxes of be and cno neutrinos directly from solar neutrino experiments , with minimal assumptions about solar models .
next we compare these results with solar models , both standard and non standard ones .
finally we discuss the expectations for borexino , both in the case of standard and non standard neutrinos . |
[ [ section ] ] in this work , we have analyzed the connection between the conformal quantum mechanics @xcite and the wigner - heisenberg algebra @xcite . with an appropriate relationship between the coupling constant @xmath64 and real constant @xmath63
one can identify the wigner hamiltonian with the conformal hamiltonian @xmath14 .
the important result is that the introduction of the wigner - heisenberg algebra in the conformal quantum mechanics is still consistent with the conformal symmetry .
the spectrum for the casimir operator , the hamiltonian @xmath14 and the ladder operators depend on the parity operator .
we also investigated the supersymmetrization of this model , in that case we obtain a new spectra for the supersymmetric and conformal hamiltonian of the calogero interaction s type . in this case the spectrum for the super - casimir operator and the superhamiltonian depend also on the parity operator .
one motivation for this work is the future applications in the problems related to black holes @xcite and the construction of a supersymmetric quantum mechanics with conformal symmetry for the n - particles in the wigner - heisenberg picture .
let us point out that one can consider the same analysis implemented in this work for the dirac oscillator , getting a generalization of the works @xcite . finally , it would be interesting to investigate the possible connection between the hidden supersymmetry in the wh - algebra structure and the explicit supersymmetry implemented in the section 4 , work in this direction is under current research .
the authors are grateful to prof .
j. a. helayel neto for the kind hospitality at cbpf - mct .
the authors are also partially supported by cnpq - claf .
we are also grateful to harold blas and f. toppan for interesting discussions and suggestions .
r. de lima rodrigues and a. n. vaidya , _ dirac oscillator via r - deformed heisenberg algebra _ , proceedings of the xxiii brazilian national meeting on particles and fields ( october/2002 ) , site www.sbfisica.org.br/eventos/enfpc/xxiii/procs/trb0013.pdf , preprintcbpf - nf-030/02 , hep - th/0301199 . | we show the natural relation between the wigner hamiltonian and the conformal hamiltonian . it is presented a model in ( super)conformal quantum mechanics with ( super)conformal symmetry in the wigner - heisenberg algebra picture @xmath0= i(1+c{\bf p})$ ] ( @xmath1 being the parity operator ) . in this context , the energy spectrum , the casimir operator , creation and annihilation operators are defined .
this superconformal hamiltonian is similar to the super - hamiltonian of the calogero model and it is also an extension of the super - hamiltonian for the dirac oscillator . where @xmath2,\ ] ] so that the parity operator is conserved , i.e. , @xmath3 = 0.$ ] when one introduces the following operators @xmath4 it can be shown that these operators together with the conformal quantum mechanics operators @xmath5 and @xmath6 @xmath7 satisfy the deformed superalgebra @xmath8 when we fix @xmath9 or @xmath10 .
] , viz . , @xmath11 & = & -2\imath \mathcal{h } , \nonumber\\ \left[{\mathcal{h}},k \right ] & = & -\imath d , \nonumber\\ \left[k , d \right ] & = & 2 \imath k,\nonumber\\ \{q_{c } , q_{c}^\dagger \ } & = & 2 \mathcal{h},\nonumber \\
\{q_{c } , s^\dagger\ } & = & -\imath d - \frac{1}{2 } b(1 + c { \bf p } ) + \sqrt{g } , \nonumber\\ \{q_{c}^\dagger , s \ } & = & + \imath d - \frac{1}{2 } b(1 + c { \bf p } ) + \sqrt{g } , \nonumber\\ \left[q_{c}^{\dagger } , d \right ] & = & -\imath q_{c}^\dagger,\nonumber\\ \left[q_{c}^\dagger , k \right ] & = & s^\dagger , \nonumber\\ \left[q_{c}^\dagger , b \right ] & = & 2 q_{c}^\dagger , \nonumber\\ \left[q_{c } , k \right ] & = & -s,\nonumber\\ \left[q_{c } , b \right ] & = & -2 q_{c } , \nonumber\\ \left[q_{c } , d \right ] & = & -\imath q_{c } , \nonumber\\ \left [ \mathcal{h } , s \right ] & = & q_{c } , \hspace{2 cm } \left [ \mathcal{h } , s^\dagger \right ] = -q_{c}^\dagger , \nonumber\\ \left [ b , s^\dagger \right ] & = & -2 s^\dagger , \hspace{2 cm } \left[b , s \right ] = 2s , \nonumber\\ \left[d , s \right ] & = & -\imath s , \hspace{2 cm } \left[d , s^\dagger \right ] = -\imath s^\dagger,\nonumber\\ \{s^\dagger , s \ } & = & 2 k,\end{aligned}\ ] ] where , @xmath12 are bosonic operators and @xmath13 are fermionic operators .
this superalgebra was introduced by plyushchay in the bosonization of supersymmetry context @xcite .
the supersymmetric extension of the conformal hamiltonian @xmath14 ( presented in the previous section ) is @xmath15 = 0.\ ] ] the super - hamiltonian @xmath16 is similar to the supersymmetric calogero interaction hamiltonian associated with two - particle interaction .
the possibility that the @xmath17-particle calogero model bosonic particles interacting through the inverse square and harmonic potential , it is completely integrable ( at classical and quantum levels ) . ] might be relevant for a microscopic description of the extreme reissner - nordstrom black hole , has been discussed in refs .
@xcite .
specifically , it was conjectured by gibbons and townsend that the large @xmath17 limit of @xmath18 superconformal version of the @xmath17-particle calogero model might provide microscopical description of the reissner - nordstrom black hole near to the horizon @xcite . continuing with our discussion
, now we are able to build up the ladder operators for the spectral resolution of the super - hamiltonian @xmath16 .
note that defining @xmath19 @xmath20 we obtain @xmath21 & = & -2 q^\dagger , \nonumber\\ \left [ h , q \right ] & = & 2 q,\end{aligned}\ ] ] and @xmath22 & = & 2 m^\dagger , \nonumber\\ \left [ \tilde{h } , m \right ] & = & -2 m .\end{aligned}\ ] ] thus we see that @xmath23 and @xmath24 and @xmath25 are ladder operators associated to the susy hamiltonian operators @xmath26 and @xmath27 , respectively .
we remark that since @xmath28 and @xmath29 commute with each other , they form a set of mutually commuting operators .
the superhamiltonian @xmath26 and @xmath27 are extensions of the ones in references @xcite for the superconformal dirac oscillator .
the supersymmetric generalizations of the ladder operators defined in ( [ ll0 ] ) in the bosonic case , are given by @xmath30 with the following commutation relations of the @xmath31 algebra @xmath32 & = & -\mathcal{l}_{+ } , \nonumber\\ \left[{\mathcal{l}}_{- } , { \mathcal{h}}_{0 } \right ] & = & + \mathcal{l}_{- } , \nonumber\\ \left[\mathcal{l}_{+ } , \mathcal{l}_{- } \right ] & = & - 2{\mathcal{h}}_{0}.\end{aligned}\ ] ] the supersymmetric casimir operator is given by @xmath33 thus , the energy spectrum becomes @xmath34 with @xmath35 .
since the eigenket @xmath36 is normalized , from ( [ cl2 ] ) and ( [ eau ] ) we obtain @xmath37 in this case , we may write the analogous representations to the ladder operators given by eq .
( [ lmas ] ) @xmath38 where @xmath39 , and @xmath40 have a similar form to the bosonic case , given in the equations ( [ lmas ] ) .
explicitly @xmath41\end{aligned}\ ] ] where @xmath42 , \hspace{3 cm } b = \left [ \begin{array}{cc } -1 & 0\\ 0 & 1 \end{array } \right]= -\sigma_{3 } , \label{const}\end{aligned}\ ] ] and the parity operator must be substituted by its eigenvalues on the eigenstates @xmath43 in the case @xmath44 on gets @xmath45,\ ] ] with the following eigenvalues @xmath46 from equation ( [ h0 ] ) and ( [ const ] ) we obtain the susy conformal hamiltonian in the coordinate representation @xmath47,\end{aligned}\ ] ] with the following superconformal potential @xmath48 = \frac 14\left((x^2+\frac{c^2 + 2c+4g}{4x^2}){\bf 1}_{2\hbox{x}2 } - \sigma_3\frac{\sqrt{g}(1-c ) } { x^2}\right).\end{aligned}\ ] ] in fig .
1 we plot @xmath49 , the bosonic part of this supersymmetric potential . in the case
@xmath50 we get @xmath51.\end{aligned}\ ] ] also , in this case @xmath52\end{aligned}\ ] ] with superconformal potential @xmath53 in fig .
2 we plot @xmath54 , the bosonic part of this supersymmetric potential . observe that we are using the convention that the operator number @xmath55 has the fermion number @xmath56 and is associated to the bosonic state and this eigenstate
is given by @xmath57 similarly , the eigenstates of @xmath58 with the fermion number @xmath59 is associated to the fermionic state , and is given by @xmath60 finally , observing the supersymmetric hamiltonians @xmath61 and @xmath62 we recall that the parameter @xmath63 is real and @xmath64 must be positive .
with appropriate choice of these parameters we can recover various cases , for example the oscillator type superhamiltonian ( without the presence of the proportional term @xmath65 ) although the bosonic part without supersymmetry has this term @xmath66 . in this case
the supersymmetric potential will not have a singularity at the origin of coordinates . |
energy harvesting communication systems involve transmitters being powered by environmental sources such as solar , vibration , and thermal effects , either alone or as supplement to the power drawn from a grid .
the ability to supply the energy storage units from environmental sources can be very useful in distributed systems such as wireless sensor networks , and m2 m networks .
recent developments in ambient energy harvesting technologies have already resulted in the practical implementation of such systems @xcite .
dependence on a variable energy source ( as opposed to a constant supply of power ) poses interesting new challenges for the transmission of information .
optimal adaptation of transmission rate has been analyzed under various problem formulations @xcite-@xcite . in @xcite ,
the transmission time mimimization problem on an awgn channel of data packets arriving at arbitrary but known time instants , using energy harvests occuring again at arbitrary , known time instants was formulated and solved . in @xcite and @xcite ,
the formulation was extended to a broadcast channel with a static data pool .
this was extended to cover the case of new data arriving during transmission in @xcite . in @xcite ,
@xcite bounds on the capacity of an energy harvesting awgn channel were obtained .
the results were extended to a fading channel in @xcite .
the solution of the transmission completion time minimization problem on a fading channel with a static data pool and known harvest times and channel states was reported in @xcite .
this paper extends the formulation of @xcite by relaxing the static data pool assumption , and its main contribution is to develop an offline solution for the time minimizing packet scheduling problem with a rechargeable transmitter under fading conditions . the solution needs to adjust its transmission power and rate over the course of transmission with respect to packet arrivals , as well as channel state and energy harvests .
this will sometimes correspond to lowering rate ( therefore the energy per bit ) , to work energy efficiently and prevent premature data queue idleness ; and at other times , increasing the rate to take advantage of a good channel state , especially when energy is abundant .
the problem statement is made precise in the next section .
after making the problem statement precise , we show the uniqueness of its solution through equivalence to a related throughput maximization problem .
the solution is first obtained for the equivalent convex problem by unconstrained sequential minimization technique ( sumt ) and the optimal allocation of the completion time minimization problem is shown to be achievable through iterative runs of sumt .
we begin with the problem statement in the next section .
consider point - to - point communication over a fading channel , where transmission is supplied by the harvested energy , arriving at arbitrary instants . following the _ offline _ formulation in @xcite
, we assume that the transmitter has knowledge of the energy harvests as well as channel states before transmission starts .
in contrast to @xcite , data packets are allowed to arrive at arbitrary ( known ) times during the course of transmission .
the harvested energy is stored in an ( ideal ) battery and immediately becomes available for use by the transmitter .
data packets are stored in a data buffer ( of infinite capacity . )
an example sequence of energy and packet arrivals , as well as channel gain changes is illustrated in fig .
[ fig : sm ] . starting from time @xmath0 ,
the amounts of energy and data have become available by time @xmath1 are denoted by @xmath2 and @xmath3 , respectively .
any arrival of energy or data or a change in the channel state is called an _ event_. the duration between any two consequent events is called an epoch .
the length of @xmath4 epoch is @xmath5 .
given an average power constraint @xmath6 and channel gain level @xmath7 during the @xmath4 epoch , we assume rate level of @xmath8 is achievable for a certain tolerable error probability . equivalently the power level used to transmit a codeword at rate @xmath9 is given by : @xmath10 .
accordingly , in discrete - time , the received signal is @xmath11 , where @xmath12 and @xmath13 are the input symbol and channel gain , and @xmath14 is zero - mean unit variance gaussian noise . given an average power constraint @xmath15 , capacity of the channel is given by @xmath16 .
, @xmath17 are event times ( energy harvests marked as @xmath18 , channel state changes , @xmath19 , or data arrivals @xmath20 ) .
the @xmath21 denote inter - event epoch durations . ]
we consider packets arriving in a certain time window of size @xmath22 .
the problem is to find an allocation of power and rate across time that minimizes the total duration of transmission for all of these packets .
an optimal policy should respect causality constraints ( at any time , only the resources available up to that point can be used ) .
it immediately follows from the concavity of the rate function that rate ( and power ) should not change within an epoch .
so , the search for an optimal schedule can be limited to schedules that keep a constant power level and rate within each epoch .
hence , the optimization problem can be written in terms of rates assigned to epochs .
note that in the problem formulation below , the solution space is further limited ( w.l.o.g . ) to schedules spanning no more than some @xmath23 epochs .
the value of @xmath23 can be set as the number of epochs used by any arbitrary feasible schedule .
[ pr : fadeschedulingtmin ] * transmission time minimization of packets on an energy harvesting fading channel : * @xmath24 in pr .
[ pr : fadeschedulingtmin ] , @xmath25 denotes the last epoch used in an optimal schedule .
state the ( energy and data ) causality constraints , while ensures transmission completion of all data .
following @xcite , we will exhibit the equivalence of pr .
[ pr : fadeschedulingtmin ] to a convex problem , namely problem [ pr : fadeschedulingdmax ] .
problem [ pr : fadeschedulingdmax ] aims to find a schedule which minimizes total energy consumption to transmit a given sequence of packets within a given deadline constraint @xmath26 , using the energy harvested during this time .
[ pr : fadeschedulingdmax ] * energy consumption minimization of packets on an energy harvesting fading channel : *
@xmath27 [ lmm : convexproblem ] problem [ pr : fadeschedulingdmax ] is a convex optimization problem .
_ firstly , note that @xmath28 is a strictly convex , monotonically increasing function .
furthermore , the constraint set of the problem is defined by non - negative weighted sums of either @xmath28 s and @xmath9 s , each constraint being either convex or linear , respectively .
it easily follows ( please see @xcite for details ) that the set of feasible allocations form a convex region .
finally , as the objective function of the minimization problem is also a non - negative weighted sum of increasing convex functions , we conclude that pr .
[ pr : fadeschedulingdmax ] is convex @xcite .
[ lmm : equivalence ] suppose @xmath26 is the minimum completion time ( obtained by solving pr .
[ pr : fadeschedulingtmin ] ) for the sequence of packets arriving by time @xmath29 , @xmath30 .
then , for this sequence of events , any solution of pr . [ pr : fadeschedulingdmax ] with deadline constraint specified as @xmath26 provides a solution to pr .
[ pr : fadeschedulingtmin ] . _ proof .
_ let schedules @xmath31 and @xmath32 be optimal solutions of pr .
[ pr : fadeschedulingtmin ] , and pr .
[ pr : fadeschedulingdmax ] , defined with the deadline @xmath26 , respectively
. the energy consumption of both schedules must be the same since the opposite claim would contradict the optimality of the schedules : @xmath32 has used no more energy by @xmath26 than @xmath31 , by definition .
suppose it used less energy , than this means that @xmath31 has used some extra energy to transmit the same packets as @xmath32 .
but then , it could use this extra energy in the last epoch to reduce the completion time by a nonzero amount , which would contradict optimality .
hence , we have 2 schedules completing the transmission of the same amount of data at the same time by consuming same amount of energy . thus , @xmath31 and @xmath32 are both solutions to problems [ pr : fadeschedulingtmin ] and [ pr : fadeschedulingdmax ] .
but by convexity , pr .
[ pr : fadeschedulingdmax ] has a unique solution , hence @xmath31 and @xmath32 must be the same . ' '' '' + _ proof . _
consider a schedule @xmath31 , which is a solution of pr .
[ pr : fadeschedulingtmin ] , and a schedule @xmath32 , which provides an optimal solution to pr .
[ pr : fadeschedulingdmax ] , defined with the deadline @xmath26 , respectively
. the energy consumption of both schedules must be the same since the opposite claim will contradict the optimality of each allocation .
now we have 2 schedules completing the transmission of the same amount of data at the same time by consuming same amount of energy .
thus , @xmath31 and @xmath32 are both solutions to problems [ pr : fadeschedulingtmin ] and [ pr : fadeschedulingdmax ] .
but by convexity , pr .
[ pr : fadeschedulingdmax ] has a unique solution , hence @xmath31 and @xmath32 are the same . ' '' '' + _ proof .
_ consider two schedules @xmath31 and @xmath32 , which respectively provide optimal solutions to problems [ pr : fadeschedulingtmin ] and [ pr : fadeschedulingdmax ] , respectively .
let @xmath33 be the completion time of @xmath31 , and the deadline constraint of @xmath32 . by time @xmath33 the energy consumption of both schedules must be the same since the opposite claim will contradict the optimality each allocation .
now we have 2 schedules completing the transmission of the same amount of data at the same time by consuming same amount of energy .
thus , @xmath31 and @xmath32 are both solutions to problems [ pr : fadeschedulingtmin ] and [ pr : fadeschedulingdmax ] .
but by convexity , pr .
[ pr : fadeschedulingdmax ] has a unique solution , hence @xmath31 and @xmath32 are the same . ' '' '' + [ corollary : uniqueness ] solution of pr .
[ pr : fadeschedulingtmin ] is unique .
_ as any solution of pr .
[ pr : fadeschedulingtmin ] provides a solution to pr .
[ pr : fadeschedulingdmax ] , in order for pr .
[ pr : fadeschedulingdmax ] to have a unique solution ( which is true by convexity ) pr .
[ pr : fadeschedulingtmin ] must have a unique solution . ' '' '' + the _ sequential unconstrained minimization technique ( sumt ) _ is a convenient method @xcite for iteratively converging to the optimal of a constrained problem by solving a sequence of unconstrained optimization problems .
the unconstrained problems are formed by adding to the objective of the original problem penalty terms corresponding to constraint violations .
correspondingly in our case , we obtain pr .
[ pr : unconstrained_minimization ] as follows : [ pr : unconstrained_minimization ] * unconstrained minimization problem : * @xmath34 @xmath35 due to constraints , and , penalty terms , , and have been added to the objective function .
starting from a point in the exterior of the feasible region for an initial value of the penalty coefficient @xmath36 , we reach the next point by solving the corresponding unconstrained minimization problem . at each sumt iteration ,
initial point is moved to the previously computed result . by iterating the penalty coefficient such that after iteration @xmath14 , @xmath37 for some growth parameter @xmath38 , we solve a sequence of unconstrained problems with monotonically increasing values of the penalty coefficient .
intuitively , this drives the points toward the feasible region . as
the penalty is zero inside the feasible region , this means that the point the algorithm stops at will be the first feasible point.it is proved in @xcite that in the case of a convex objective and penalty terms as defined above , the algorithm converges to the optimum of the original constrained problem as @xmath39 goes to infinity .
in practice , the iterations are stopped when an arbitrary stopping criterion @xmath40 is satisfied . in our problem
, sumt is initialized with an infeasible allocation ( i.e. , at a point in the exterior of the constraint region ) , specifically , transmitting all data at constant rate within the given deadline @xmath26 , disregarding causality constraints . to ensure fast convergence ( see @xcite ) @xmath39
is initialized such that the values of the objective and penalty terms are commensurate , and the penalty terms corresponding to each constraint are scaled such that no constraint dominates . at each iteration of sumt
, the corresponding unconstrained problem is solved by _
newton s method_. it is quite standard to apply newton s method in the inner iterations of sumt . after the @xmath41 newton step in an inner iteration , rate allocation vector is updated as : @xmath42^{-1 } \nabla f(\textbf{r}^l)$ ] .
the _ newton decrement _ , @xmath43^{-1 } \nabla f(\textbf{r}^l)\right)^{\frac{1}{2}}$ ] , becoming smaller than a predefined accuracy parameter @xmath44 is the stopping criterion for each inner iteration . by reducing @xmath44 ,
the inner optimizations can be made arbitrarily accurate @xcite .
the convergence rate of these iterations will be discussed in the following sections . from lemma 2 , using the optimal value of completion time , @xmath45 , as a parameter in pr.2 would give us an optimal schedule for pr.1 .
of course , @xmath45 is not known before solving pr.1 .
the method we will use is to iteratively approach @xmath45 by solving pr . 2 for different values of @xmath26 and checking the resulting amount of energy consumption .
the bisection method will be used to monotonically narrow down the interval in which the optimal completion time @xmath46 of pr .
[ pr : fadeschedulingtmin ] must lie in . since @xmath47 is a monotonically decreasing and continuous function of @xmath26 @xcite , any feasible value of @xmath26 provides an upper bound on @xmath46 . in search of upper and lower bounds
, @xmath26 is initialized as the end of last data arrival epoch , and sumt is run as detailed in section a. if the resulting optimal energy that sumt returns is too high , it means that transmission can not be completed within this deadline , hence the current value of @xmath26 provides a lower bound .
@xmath26 is then is extended by the next epoch length .
this procedure is repeated until the total consumed energy returned by sumt goes below the energy harvested by @xmath26 .
that value of @xmath26 provides an upper bound .
the next deadline is chosen as the average of the upper and lower bounds , and sumt is run again .
if the deadline is feasible , it becomes the new upper bound , if not , it becomes the new lower bound , and so on .
the iterations are stopped when the difference between the upper and lower bounds goes below @xmath48 , which , provided that the inner optimizations of sumt are also done with sufficient accuracy , sandwiches @xmath46 in an interval of size @xmath48 .
the computational complexity is largely imposed by the stopping criteria of newton , sumt and bisection iterations . to compute the overall complexity of proposed scheme ,
let us first consider the number of bisections .
when bisection iterations begin , the difference between upper and lower bounds on completion time becomes the last epoch length of the most current schedule returned by sumt . in each iteration this interval is halved , so at most @xmath49 bisections are to be performed .
for each bisection , sumt makes @xmath50 iterations to converge with a desired accuracy of @xmath51 @xcite .
the number of newton steps to achieve an accuracy of @xmath44 in the inner newton iterations per each iteration of sumt is upper bounded by @xmath52 , where @xmath53 is the minimum decrement amount of @xmath54 and @xmath55 is the value at the optimal point @xcite .
this bound follows from the different nature of convergence of newton s iterations for different operating points .
it has been shown in @xcite that , once the operation point gets sufficiently close to the optimum , convergence rate is quadratic , while it is approximately linear until then . finally , the computational requirements imposed by each newton step , due to the construction and the inversion of a @xmath56 hessian ( where @xmath57 is the number of epochs in the problem ) , is polynomial ( with complexity @xmath58 or as low as @xmath59 with ultimately efficient implementation . )
as an example , we consider the event sequence depicted in fig .
[ fig : numex ] . the final schedule returned by the proposed algorithm
is also shown in the figure .
the penalty parameter @xmath39 is initialized as @xmath60 , the growth parameter @xmath61 is set to @xmath62 .
the threshold values for newton s method , sumt and bisection are @xmath63 , @xmath64 and @xmath65 , respectively .
the algorithm repeats @xmath66 sumt iterations , within which at most 6 newton s steps are repeated , for each of @xmath67 bisection repetitions and terminates within @xmath68 seconds in matlab running on a macbook pro rev .
when the newton s and bisection thresholds are raised to @xmath65 and @xmath69 , respectively , the run time reduces to @xmath70 seconds .
we believe that optimizing the code over an efficient programming platform can reduce this time significantly .
epoch is @xmath19 , the bandwidth is @xmath71 khz , energy harvest amounts and arriving data are marked as @xmath18 and @xmath20 , respectively .
( b ) final schedule returned by completion time minimization algorithm . ]
a method for solving the offline minimum completion time packet scheduling problem on an energy harvesting fading channel has been developed and demonstrated .
the key to the method is exhibiting equivalence to an energy minimization problem which is a convex program . in certain realistic scenarios ,
the harvest profile and data arrivals may be known in advance . in that case
, the offline solution would apply for a static channel . on a fading channel with an ergodic channel state process ,
an online algorithm such as waterfilling could run on top of the offline adaptation .
when the data and/or harvest arrivals are also unknown , the offline solution here may be combined with a prediction or learning scheme or a simple look - ahead policy .
bib k. lin , j. hsu , s. zahedi , d. lee , j. friedman , a. kansal , v. raghunathan , m. srivastava , heliomote : enabling long - lived sensor networks through solar energy harvesting , " in _ proceedings of the 3rd international conference on embedded networked sensor systems ( sensys ) _ , san diego , ca , usa , p. 309
, 2005 .
j. yang and s. ulukus , optimal packet scheduling in an energy harvesting communication system , " _ ieee trans . on communications _
, 60(1):220 - 230 , january 2012 .
m. a. antepli , e. uysal - biyikoglu , and h. erkal , optimal packet scheduling on an energy harvesting broadcast link , " to appear on _ ieee journal on selected areas in communications ( jsac ) , special issue on energy - efficient wireless communications _ , 2011 .
j. yang , o. ozel and s. ulukus , broadcasting with an energy harvesting rechargeable transmitter , " _ ieee trans . on wireless communications _ , vol.pp , no.99 , pp.1 - 13 , dec .
f. m. ozcelik , h. erkal and e. uysal - biyikoglu , optimal offline packet scheduling on an energy harvesting broadcast link , " _ 2011 ieee int .
symposium on information theory _ ,
st.petersburg , aug .
h. erkal , f. m. ozcelik and e. uysal - biyikoglu , optimal offline packet scheduling on an energy harvesting broadcast link,"arxiv:1111.6502v1 , 2011 o. ozel and s. ulukus , information - theoretic analysis of an energy harvesting communication system , " in _ ieee pimrc workshops _ , 2010 .
r. rajesh , v. sharma , and p. viswanath , information capacity of energy harvesting sensor nodes , " _ 2011 ieee int .
symposium on information theory _ ,
st.petersburg , aug .
r. rajesh , v. sharma and p. viswanath , capacity of fading gaussian channel with an energy harvesting sensor node , " in _ ieee globecom11 , houston _ , texas , dec .
o. ozel , k. tutuncuoglu , j. yang , s. ulukus and a. yener , transmission with energy harvesting nodes in fading wireless channels : optimal policies , " _ ieee jour . on
selected areas in communications , 29(8):1732 - 1743 _ , september 2011 .
h. erkal , _ optimization of energy harvesting wireless communication systems _ , m.sc .
thesis , middle east technical university , dept . of electrical and electronics engineering , ankara , turkey , dec .
2011 , available at http://www.eee.metu.edu.tr/~cng/. f. m. ozcelik , _ optimal and implementable transmission schemes for energy harvesting networks _ , m.sc .
thesis , metu , ankara , turkey , expected 2012 , working copy available at http://www.eee.metu.edu.tr/~cng/ , last updated on june 2012 .
s. boyd and l. vandenberghe , _ convex optimization _ , cambridge university press , 2004 .
d. g. luenberger and y. ye , _ linear and nonlinear programming _ , new york : springer , 2008 .
p.a.jensen and j.f .
bard , _ algorithms for constrained optimization _ , 2003 , available at http://www.me.utexas.edu/~jensen/ormm/supplements/units/nlp_methods/const_opt.pdf .
consider two feasible allocation vectors @xmath72 and @xmath73 and let @xmath74 be a linear combination of these .
we will show that rate allocation vector @xmath75 is also feasible and consumes an amount of energy less than @xmath76 .
first , let us compute the energy consumption of @xmath74 .
@xmath77 follows from the strict convexity of @xmath78 function and equality holds only if @xmath79 or @xmath80 . with this inequality ,
@xmath74 allocation consumes less than @xmath76 .
we are now ready to check feasibility of the allocation and begin with the energy causality constraint , .
@xmath81 states the satisfaction of energy causality constraint .
similarly , @xmath74 also respects data causality , .
as @xmath72 and @xmath73 are two feasible schedules , implies that @xmath74 satisfies data causality constraint .
in addition , we have the following when the transmission ends : @xmath83 shows the satisfaction of .
so , we have shown that feasible allocations form a convex region , , any linear combination of two feasible schedules is also feasible . combining this result with
, we conclude that pr .
[ pr : fadeschedulingdmax ] is a convex optimization problem . | the offline problem of transmission completion time minimization for an energy harvesting transmitter under fading is extended to allow packet arrivals during transmission .
a method for computing an optimal power and rate allocation ( i.e. , an optimal offline schedule ) is developed and studied .
transmission of packets over a fading channel under energy harvesting energy harvesting , completion time , offline schedule , packet scheduling , causality constraints , energy constraint , unconstrained problem , sumt , sequential optimization , complexity . |
the interplay between disorder and superconductivity has attracted sustained interest over the past few decades . in bcs superconductors , according to anderson s theorem , @xcite the superconducting critical temperature @xmath1 and the energy gap @xmath2 remain unaffected by the presence of weak disorder , although in unconventional superconductors both @xmath1 and @xmath2 are strongly suppressed with weak disorder .
@xcite in the strongly disordered regime , things become interesting even in bcs superconductors : in two - dimensional ( 2d ) systems , strong disorder leads to a universal superconductor - to - insulator transition at a critical sheet resistance of @xmath3 ; @xcite furthermore , it has been found that , as @xmath4 tends to zero , there emerges a broad range of magnetic field where the resistivity remains finite but is much smaller than the normal - state value , which arguably signifies the importance of quantum phase fluctuations .
@xcite in three - dimensional ( 3d ) systems , strong disorder leads to a disruption of the global superconducting coherence through reductions in both the superfluid density @xmath5 and pairing interactions , which results in the superconducting order parameter to spatially fluctuate and gives rise to anomalous electronic states near the superconductor - to - metal transition .
@xcite experimentally , while the 2d systems have been actively studied in the past , @xcite the role of strong disorder in 3d superconductors is just beginning to be addressed with a modern viewpoint , @xcite which naturally calls for explorations of suitable materials for such a study . in this paper , we report that agsnse@xmath6 , which is a low-@xmath7 conventional superconductor having a cubic structure , offers an ideal playground for studying the effects of strong disorder in 3d superconductors .
we show that one can synthesize high - quality samples of agsnse@xmath6 presenting a very sharp superconducting transition , which indicates that morphologically the system is homogeneous ; nonetheless , strong electron scatterings that are intrinsic to this system lead to an extremely type - ii superconductivity with the ginzburg - landau parameter @xmath8 of as large as 55 .
this causes the emergence of a broad range of magnetic fields where the resistivity remains finite but is much smaller than the normal - state value as @xmath9 , which is very much reminiscent of the behavior observed in disordered 2d superconductors @xcite or in quasi-2d systems like high-@xmath7 cuprates .
@xcite furthermore , such an anomalous metallic state is found in the magnetic field range _ above _ the upper critical field determined thermodynamically from the specific heat .
since this anomalous state can be easily reached with the magnetic field of 3 t and the microscopic mechanism of superconductivity is well understood in conventional superconductors , future studies of agsnse@xmath6 would help elucidate the roles of quantum phase fluctuations which remain controversial to date .
the superconductivity in agsnse@xmath0 was discovered by johnston and adrian @xcite in 1977 as a possible valence - skipping superconductor . in valence - skipping compounds ,
the nominal valence state of a constituent element happens to be the skipped valence for that element .
this characteristic has been of particular interest for superconductivity , since it may provide a means to realize the so - called negative-@xmath11 mechanism to enhance the superconducting transition temperature @xmath7 . @xcite
well - known examples of such valence - skipping superconductors are babi@xmath12pb@xmath13o@xmath14 , @xcite ba@xmath12k@xmath13bio@xmath14 , @xcite and pb@xmath12tl@xmath13te .
@xcite in all these compounds , @xmath7 is relatively high for their low carrier density . at the same time ,
valence fluctuations lead to significant electron scattering , making those superconductors highly disordered even though the matrix is homogeneous .
agsnse@xmath0 adopts the face - centered cubic structure , in which the cation sites are randomly occupied by equal amount of ag and sn atoms .
since the nominal valence of sn is 3 + which coincides with its skipped valence state , the sn ions should be separated into 1:1 mixture of sn@xmath15 and sn@xmath16 and the chemical formula of this material can be more appropriately expressed as ( ag@xmath17)(sn@xmath15)@xmath18(sn@xmath16)@xmath18(se@xmath19)@xmath0 , as evidenced by the measurements of the magnetic susceptibility @xcite and the @xmath20sn mssbauer spectra .
@xcite despite the valence skipping nature and the carrier density of the order of 10@xmath21 @xmath22 , its @xmath7 of 4.7 k is not particularly high , which suggests that the negative-@xmath11 mechanism is not at work here .
therefore , in this compound one can explore manifestations of strong disorder associated with valence fluctuations in the context of conventional superconductivity without being bothered by strong electron interactions . .
note that ag and sn atoms are randomly distributed in the cation sites .
( b ) xrd pattern of the agsnse@xmath0 sample .
all the diffraction peaks can be well indexed using a cubic cell with the f__m__3__m _ _ space group .
( c ) temperature dependence of @xmath23 of the agsnse@xmath0 sample .
lower inset shows an enlarged view of the data near the superconducting transition for @xmath23 ( right axis ) and for the dc magnetic susceptibility measured upon zero - field - cooling ( left axis , shown in terms of the shielding fraction ) .
upper inset shows the magnetic - field dependence of @xmath24 at 5 k ; its slope gives @xmath25 = 2.0 @xmath26 10@xmath21 @xmath22 .
, width=328 ]
we synthesized high - quality ingots of polycrystalline agsnse@xmath0 by using a two - step method as described in ref . .
first , high - purity shots of ag ( 99.999% ) , sn ( 99.99% ) , and se ( 99.999% ) with the stoichiometric ratio of 1:1:2 were melted in sealed evacuated quartz tube at 800 @xmath27c for 48 h with intermittent shaking to ensure homogeneity , followed by quenching to room temperature .
the quartz tube containing the melt - quenched ingot was subsequently annealed at 450 @xmath27c for two weeks , and then quenched into cold water .
the structure of the resulting sample was characterized by powder x - ray diffraction ( xrd ) . as can be seen in fig .
1(b ) , the xrd pattern shows sharp diffraction peaks consistent with the face - centered cubic structure , giving the lattice parameter @xmath28 of 5.675(1 ) which agrees with the literature .
@xcite the ingot was cut into bar - shaped samples , and the temperature - dependent magnetization @xmath29 under various magnetic fields was measured with a vibrating sample magnetometer ( vsm ) , while the magnetization curves at fixed temperatures were measured with a commercial squid magnetometer ( quantum design mpms- 1 ) .
the demagnetization effect was corrected for by considering the slab geometry .
@xcite the resistivity @xmath23 and the hall resistivity @xmath24 were measured by using a standard six - probe method where the contacts were made by spot - welding gold wires .
the specific heat @xmath30 was measured with a relaxation - time method using quantum design ppms-9 . for consistency reasons
, all the data presented here were measured on exactly the same sample .
figure 1(c ) shows the temperature dependence of @xmath23 , which exhibits only a weak metallic temperature dependence . this is qualitatively similar to that
reported @xcite for ag@xmath31sn@xmath32se@xmath0 and signifies the presence of intrinsically strong disorder that does not change much with the ag / sn ratio . as shown in the upper inset of fig .
1(c ) , @xmath24 is linear in @xmath33 and its slope at 5 k corresponds to an electron density @xmath34 of 2.0 @xmath26 10@xmath21 @xmath22 .
the superconducting transition in @xmath23 occurs sharply between 4.5 and 4.6 k , and the magnetic susceptibility shows the onset @xmath7 of 4.55 k , which corresponds to the midpoint of the resistive transition .
the superconducting shielding faction achieves nearly 100% after the demagnetization correction .
there is no resistivity upturn at low temperatures , giving no clear sign of anderson localization .
measured in 0 and 9 t. inset shows @xmath35 plotted as a function of @xmath36 for the 9-t data , where the superconductivity is completely suppressed ; the solid line presents the best fit to the data using the debye model ( see ref . ) .
( b ) temperature dependence of @xmath37 in 0 t. the dash - dotted line is the calculated @xmath37 curve given by the weak - coupling bcs theory ; the solid line is the best fit using a modified bcs theory with @xmath38 = 1.925 .
the horizontal solid line denotes @xmath39 .
, width=328 ] figure 2(a ) shows the temperature dependences of @xmath35 measured under 0 and 9 t. application of a 9 t field completely suppresses the superconductivity , enabling us to determine the phononic contribution and the normal - state parameters such as the electronic specific - heat coefficient @xmath39 and the effective mass @xmath40 .
@xcite figure 2(b ) shows the temperature dependence of the electronic specific heat @xmath37 in 0 t obtained by subtracting the phononic contribution determined from the 9-t data .
one can see that @xmath37 quickly approaches zero at low @xmath4 , indicating a fully - gapped nature . as for the precise temperature dependence , although the simple weak - coupling bcs model ( dash - dotted line ) fails to describe the @xmath37 data , a modified bcs model @xcite that allows the coupling constant @xmath41 to vary [ @xmath42 is the gap at 0 k ] can fit our data well with @xmath38 = 1.925 ( solid line ) , which is only slightly larger than the weak - coupling bcs value of 1.764 . using @xmath40 and @xmath38 obtained from @xmath43 , @xcite
the pippard coherence length @xmath44 can be estimated as @xmath44 = @xmath45 = 181 nm .
this is to be contrasted with the mean free path @xmath46 = @xmath47 = 0.9 nm , where @xmath48 is the residual resistivity .
this @xmath46 is less than @xmath49 and is very short , pointing to the existence of strong scattering which is consistent with the valence fluctuations of sn .
the ratio @xmath50/@xmath44 = 0.005 means that agsnse@xmath0 is in the very dirty limit . yet
, @xmath51 is estimated to be @xmath528 and hence the system is still a good metal . note that the estimations of @xmath40 , @xmath53 , @xmath44 , and @xmath46 in this paper are based on the free electron theory ; once the band structure of agsnse@xmath0 is known , those estimations can be made more accurate .
curves measured after zero - field cooling to various temperatures .
the dashed line marks the initial linear behavior .
( b ) plot of @xmath54 vs @xmath4 ; the solid line is a fit to the data using the local dirty limit formula . the @xmath55(0 ) obtained after the demagnetization correction is marked by an arrow .
inset shows the @xmath33 dependences of @xmath56 @xmath57 @xmath58 , where @xmath59 denotes the initial meissner contribution ; the data are shifted vertically for clarity . the determined @xmath60 value for each temperature is shown by arrows .
( c ) temperature dependences of @xmath61 under various magnetic fields .
the @xmath62 values are indicated by arrows .
( d ) @xmath63 vs @xmath4 phase diagram ; the solid line is a whh fit to the data .
, width=328 ] figure 3 summarizes the results of the magnetization measurements to extract the values of the upper and lower critical fields .
figure 3(a ) shows @xmath64 curves measured after zero - field cooling to various temperatures .
we define @xmath54 at each temperature as the value at which the @xmath64 data deviates from its initial linear behavior , as can be more clearly seen in the inset of fig .
. extrapolation @xcite of the resulting @xmath65 data to @xmath4 = 0 k yields @xmath66 = 1.18 mt which , after correcting for the demagnetization effect , @xcite gives the lower critical field at @xmath4 = 0 k , @xmath67 , of 1.58 mt .
figure 3(c ) shows the temperature dependences of the magnetization near the superconducting transition under various magnetic fields .
a paramagnetic background with a temperature - independent susceptibility @xmath68 = 4.2 @xmath26 10@xmath69 emu / g has been subtracted from the data , @xcite and what is plotted is @xmath70
. a gradual suppression of superconductivity with increasing magnetic field is evident . for each magnetic field
, we determined the onset temperature of the transition , @xmath62 , from the intersection of the linear extrapolation of the @xmath71 data to the @xmath72 = 0 line , as indicated by the arrows in fig .
the obtained @xmath62 data shown in fig .
3(d ) give the temperature dependence of the magnetically - determined upper critical field , @xmath73 , which is extrapolated to @xmath4 = 0 k using the werthamer - helfand - hohenberg ( whh ) theory @xcite to yield @xmath74 = 2.26 t. vs @xmath4 phase diagram ; the solid line is a whh fit to the data .
for comparison , @xmath73 and its whh fitting are shown with filled squares and a dashed line , respectively . , width=226 ] to evaluate the upper critical field at low temperatures in a more direct way , we measured the superconducting transition in specific heat under various magnetic fields .
figure 4(a ) shows representative data of @xmath37 vs @xmath4 measured in 0 , 0.25 , 1.0 , and 1.5 t , where the superconducting transition is well resolved . in higher magnetic fields ,
the specific - heat anomaly becomes weaker , as shown in fig .
4(b ) for 1.75 , 1.9 , and 2.0 t ; here , the transition can still be seen in 1.9 t , but no clear anomaly is discernible at 2.0 t. these data demonstrate that the upper critical field is thermodynamically well - defined in agsnse@xmath0 at low temperature and that it lies around 2 t near @xmath4 = 0 k. to quantitatively analyze the superconducting transition in the specific - heat data , for each magnetic field we determined three characteristic temperatures : onset of the transition ( below which the specific heat grows ) , midpoint of the transition , and the peak in @xmath75 .
it is useful to note that the transition width @xmath76 , which we define by the difference between the onset and peak temperatures , remains reasonably narrow with @xmath77 k at all temperatures .
we choose to use the midpoint of the specific - heat transition for defining the transition temperature @xmath78 for each magnetic field , which in turn gives the temperature dependence of the upper critical field @xmath79 .
the solid line in fig .
4(c ) is the fitting of the whh theory to the @xmath79 data , which gives the @xmath4 = 0 k value @xmath80 of 2.11 t. for comparison , we reproduced in fig .
4(c ) the @xmath81 data obtained from the magnetization measurements .
the difference in the whh fittings for the specific - heat and magnetization data is only 7% , which indicates that the two thermodynamical measurements are essentially consistent . in the following discussions , we take @xmath82 as the thermodynamically - determined @xmath83
. now we estimate various superconducting parameters of agsnse@xmath0 by using @xmath84 as the zero - temperature upper critical field @xmath85 .
note that since agsnse@xmath0 has cubic symmetry , no anisotropy is expected .
the ginzburg - landau ( gl ) coherence length @xmath86 is estimated as @xmath87(0 ) = @xmath88 = 12.5 nm , where @xmath89 = 2.07
@xmath26 10@xmath90 wb is the flux quantum .
alternatively , one may also estimate @xmath87 from @xmath86 @xmath91 @xmath92 for dirty superconductors , which yields @xmath87 @xmath91 12.7 nm .
the consistency is gratifying .
the equation @xcite @xmath55(0)/@xmath93(0 ) = @xmath94 gives the gl parameter @xmath95 = 55 , which in turn allows us to calculate the effective penetration depth @xmath96 = 685 nm through @xmath97 = @xmath98 .
this gives the london penetration depth @xmath99 @xmath91 @xmath100 = 48.3 nm . since agsnse@xmath0 is a very dirty superconductor , the superfluid density @xmath101 is expected to be reduced by a factor of @xmath52 @xmath102 .
@xcite indeed , the calculation of @xmath101 based on @xmath101 = @xmath103 yields @xmath101 @xmath104 9 @xmath26 10@xmath105 @xmath22 , which is close to the estimation of @xmath101 @xmath104 @xmath106 = 1 @xmath26 10@xmath107 @xmath22 , reassuring the applicability of the bcs theory to agsnse@xmath0 and the soundness of our @xmath108 measurements which can be adversely affected by flux pinning . that is used for determining @xmath109 , @xmath110 , and @xmath111 are indicated by dashed lines . ,
despite the conventional nature of the superconductivity in agsnse@xmath0 , nontrivial physics becomes evident when one looks at the magnetic - field - induced resistive transition at fixed temperatures ( fig .
one can easily see that the resistive transition is very sharp at temperatures close to @xmath7 , but it becomes increasingly broader at lower temperatures .
we determined the values of the magnetic field at which @xmath23 recovers a certain percentage of the normal - state resistivity @xmath112 for each temperature , and denote them @xmath113 , @xmath114 , _ etc .
_ we also determined the irreversibility field @xmath115 to mark the onset of a finite resistivity within our sensitivity limit , which corresponds to 0.15% of @xmath112 .
the results are shown in fig .
6 , where we also plotted the thermodynamically - determined @xmath116 with a solid line .
usually , either @xmath114 or @xmath117 are considered to give a measure of the upper critical field .
@xcite in the present case , for temperatures down to @xmath118 , @xmath114 agrees reasonably well with the thermodynamically determined @xmath83 , but at lower temperatures the resistive transition occurs largely above @xmath83 .
this means that at low temperature , there is a wide range of magnetic field above @xmath83 where the resistivity is finite but is noticeably smaller than @xmath112 .
the existence of such an anomalous metallic state above @xmath83 has never been documented for a conventional 3d superconductor , and this is the main finding of the present work .
note that the pauli paramagnetic limit is estimated to be 8.4 t , which is well above the anomalous metallic regime observed here .
determined thermodynamically from specific - heat measurements ( solid line ) together with @xmath119 , @xmath109 , @xmath110 , and @xmath111 obtained from the resistive transition data . , width=245 ] similar broadening of the magnetic - field - induced resistive transition at low temperatures has been observed in high - temperature superconductors such as bi@xmath6sr@xmath6cuo@xmath120 , @xcite tl@xmath6ba@xmath6cuo@xmath120 , @xcite mgb@xmath6 , @xcite and ba@xmath12k@xmath121bio@xmath122 .
@xcite this phenomenon is generally believed to be caused by enhanced vortex fluctuations and the resulting vortex liquid phase due to a short coherence length that is inherent to high - temperature superconductors . in light of this common belief ,
the present observation in agsnse@xmath0 is striking . here
the coherence length is not so short , 12.1 nm , and the system is 3d , both of which work against strong vortex fluctuations .
therefore , our result calls for a reinterpretation of the low - temperature broadening of the magnetic - field - induced resistive transition . to understand the above results ,
an interesting possibility is the scenario proposed by spivak , oreto and kivelson @xcite that the strong disorder leads to formations of superconducting puddles " that are josephson - coupled to each other , and quantum fluctuations of the relative phase between those puddles give rise to a broad range of magnetic fields where the resistivity remains finite and smaller than @xmath112 as @xmath9 .
if this is indeed the case , our data give evidence for the important role of quantum phase fluctuations in disordered 3d conventional superconductors .
since the superconductivity mechanism is well understood here ( as opposed to the case of high - temperature superconductors ) and the required magnetic field to access the quantum fluctuation regime is low ( only @xmath523 t ) , agsnse@xmath0 provides a convenient platform for studying the role of quantum fluctuations in the magnetic - field - induced superconductor - to - metal transition .
another anomalous behavior observed here is that at low temperatures , not only @xmath111 but also @xmath119 show little tendency toward saturation .
a similar behavior was previously observed in strongly - disordered amorphous superconductors @xcite and was discussed to be due to the weak - localization effect .
@xcite however , the discrepancy between the resistive transition and the thermodynamically - determined @xmath83 makes the conclusion of those old studies @xcite questionable .
hence , the physical meaning of the resistive transition in disordered superconductors had better be reconsidered by taking into account the important role of quantum phase fluctuations .
we have fully characterized the agsnse@xmath0 superconductor where its valence - skipping nature leads to intrinsically strong electron scattering without any obvious enhancement of @xmath7 . at low temperature , we observed an anomalous broadening of the magnetic - field - induced resistive transition which occurs largely above the thermodynamically - determined @xmath83 .
this means the existence of an anomalous metallic state above @xmath83 where quantum phase fluctuations associated with a magnetic - field - induced superconductor - to - metal transition are likely to be playing an important role .
this makes agsnse@xmath0 a useful platform for studying the role of quantum fluctuations in 3d disordered superconductors . as can be seen in the inset of fig .
2(a ) , the 9 t data below 5 k can be well described by the debye model @xmath35 = @xmath39 + @xmath123@xmath36 + @xmath124 , where @xmath39 is the normal - state electronic specific - heat coefficient and @xmath125 ( @xmath126 = 3 , 5 ) denote the coefficients of the phononic contribution .
the obtained parameters are @xmath39 = 3.5 mj / molk@xmath127 , @xmath123 = 0.7 mj / molk@xmath128 , and @xmath129 = 0.009 mj / molk@xmath130 .
the @xmath123 value gives the debye temperature @xmath131 = 220 k via @xmath131 = ( 12@xmath132@xmath128@xmath133/5@xmath123)@xmath134 with @xmath135 = 4 and @xmath136 = 8.314 j / molk .
in addition , if one assumes a single spherical fermi surface , the effective mass @xmath40 can be estimated as @xmath40 = @xmath137 @xmath91 1.5@xmath138 , where @xmath139 = 55 @xmath140/mol and @xmath138 is the free electron mass . | we report superconducting properties of agsnse@xmath0 which is a conventional type - ii superconductor in the very dirty limit due to intrinsically strong electron scatterings .
while this material is an isotropic three - dimensional ( 3d ) superconductor with a not - so - short coherence length where strong vortex fluctuations are _ not _ expected , we found that the magnetic - field - induced resistive transition at fixed temperatures becomes increasingly broader toward zero temperature and , surprisingly , that this broadened transition is taking place largely _ above _ the upper critical field determined thermodynamically from the specific heat .
this result points to the existence of an anomalous metallic state possibly caused by quantum phase fluctuations in a strongly - disordered 3d superconductor . |
the kuramoto - sivashinsky equation ( kse ) in its dimensionless form for some field @xmath0 can be written as @xcite @xmath1 this equation stands as a paradigmatic model for chaotic spatially extended systems and can be used to study the connections between chaotic dynamics at small scales and apparent stochastic behavior at large scales .
it is an example of an extended , deterministic dynamical system that exhibits complex spatio - temporal phenomena .
the kse has been derived for the purpose of describing the intrinsic instabilities in laminar flame fronts @xcite and phase - dynamics in reaction - diffusion systems @xcite .
the equation ( [ eq : kse ] ) in one- and two - dimensional cases has been a subject of active research for about three decades , and its scaling properties have even been an object of some controversy @xcite .
this paper presents some results obtained from a less researched _ generalized _ version of kse .
there have been many different generalizations of the kse used for different purposes .
some of them involve adding damping terms to ( [ eq : kse ] ) @xcite , some introduce spatial anisotropy @xcite or additive random noise @xcite . in the case
presented here , there is an additional nonlinear term @xmath2 introduced to the kse ( [ eq : kse ] ) .
the two - dimensional generalized kuramoto - sivashinsky equation of this form with an additive gaussian white noise has been used as a model equation for amorphous solid surface growth @xcite .
the equation in this model originally has five parameters that are needed in order to reproduce the experimental data in the simulations and to examine the correspondence with microscopic properties of the surface growth process @xcite .
however , for theoretical investigations of the long - time and large - scale behavior of the system , the noise term can be neglected and the remaining deterministic equation can be rescaled into the dimensionless form with only one independent parameter @xmath3 : @xmath4 equation ( [ eq : gkse ] ) has also been used as a model for nano - scale pattern formation induced by ion beam sputtering .
the two - dimensional generalized kuramoto - sivashinsky equation ( gkse ) ( [ eq : gkse ] ) in the context of this paper describes the evolution of a ( 2 + 1)-dimensional surface , i.e. , a surface which is defined as a function on a two - dimensional plane and is growing in the direction perpendicular to that plane .
the surface profile @xmath5 is defined as the surface height @xmath0 at the position @xmath6 on the square @xmath7 ^ 2 $ ] of size @xmath8 in the plane @xmath9 at time @xmath10 or , more generally , as a function : @xmath11 ^ 2 \times \mathbb{r}^{+}\rightarrow\mathbb{r}\,.\ ] ] we solve the equation ( [ eq : gkse ] ) for different values of the parameter @xmath3 using the _ finite difference _
method with periodic boundary conditions , the time step @xmath12 , spatial discretization step @xmath13 .
such a seemingly bizarre number for the discretization step @xmath14 is actually a good approximation of the value that is needed in order for the system with periodic boundary conditions to be able to contain ordered patterns that appear in some other versions of the generalized kse .
the equation is solved for system sizes @xmath8 ranging from about 45 to about 1000 ( i.e. , on the @xmath15 lattices with @xmath16 from 63 to 1400 , where @xmath17 ) and in two cases up to about 1422 ( @xmath18 ) .
the methods of numerical solution for ( [ eq : gkse ] ) are presented and compared in @xcite .
the surface _ roughness _ @xmath19 , also called the _ surface width _ , is one of the most important characteristics of a surface @xcite .
it is defined as the standard deviation , or , synonymously , root mean square ( rms ) deviation , of the surface height @xmath5 from its average value @xmath20 , at some time @xmath10 : @xmath21 here and throughout the whole paper we denote the averaging by @xmath22 with some subscript that shows over what entities the averaging takes place .
@xmath23 denotes spatial average over the whole surface , @xmath24 temporal average , @xmath25 ensemble average , @xmath26 spatial average over all @xmath27 whose length is @xmath28 etc . for smaller values of parameter @xmath3 ( say , @xmath29 ) the kinetics of @xmath19 due to the evolution of the surface resulting from ( [ eq : gkse ] )
seems to follow a distinct pattern .
starting from a random surface with some small initial roughness @xmath30 , the roughness begins to grow almost exponentially , but at some time @xmath31 this growth slows down and later on crosses over to a stationary regime where it oscillates about its average ( _ saturation _ ) value @xmath32 ( see fig .
[ fig : w - saturation ] ) . for short times ( @xmath33 ) for surfaces with initial roughness @xmath34 evolving according to ( [ eq : gkse ] ) with parameter values @xmath35 ( _ black , dashed line _ ) , @xmath36 ( _ red , continuous line _ ) , @xmath37 ( _ blue , dotted line _ ) .
system size @xmath38 ( in lattice units , @xmath39).,scaledwidth=50.0% ] in the saturation regime , the dynamics of @xmath19 becomes a statistically stationary process with time independent average and other statistical characteristics that are the same for different realizations ( different realizations differ in the initial surface profile , as the evolution equation itself is deterministic ) .
however , in order to consider the long time behavior ( after saturation ) of a statistically stationary process , it is important to take time averages over sufficiently long time intervals , having in mind that the minimal required averaging time must be at least several times longer than the _ correlation time _ of the process , defined as the time lag value at which the normalized autocorrelation function of @xmath19 effectively falls to zero .
this correlation time might strongly depend on the parameter @xmath3 and the system size @xmath8 . in the stationary regime
the roughness is chaotically oscillating about some average value which we denote @xmath32 and call _ saturated roughness_. this value is calculated as the time average of @xmath19 in the stationary regime and is virtually the same when averaged over different time intervals of the stationary regime ( given that these intervals are sufficiently long ) and for different realizations .
this saturated surface roughness can be theoretically defined as a time average over an time interval of length @xmath40 that goes to infinity , starting from the time @xmath41 where the saturation regime is surely reached : @xmath42 our numerical investigation shows that , for system sizes @xmath16 and parameter values @xmath3 considered in this paper , it is sufficient to take @xmath43 and @xmath44 to get the saturation values @xmath32 that differ less than @xmath45 for different realizations . for most of the results presented here ,
we have used the values @xmath32 obtained using @xmath46 and @xmath47 and , furthermore , we used the averaged values of several ( from 3 to 10 ) realizations : @xmath48}\big\rangle_{\eta}\ , .
\label{w - estimate}\ ] ] in this paper we are interested in the surface patterns produced by gkse ( [ eq : gkse ] ) and the dependence of the saturated surface roughness on parameter @xmath3 and system size @xmath8 .
the surfaces , generated by eq .
( [ eq : gkse ] ) in the stationary regime have a disordered cellular structure with cells whose sizes are in a quite narrow interval ( see figs . [
fig : surf_smallscale][fig : surf1200 - 02 ] ) .
the resulting patterns seem to have similar appearance for different system sizes @xmath8 ( for sizes that are at least several times larger than the typical cell diameter ) . the usual way to investigate the surface patterns is by calculating the surface height correlation function @xmath49 , which is the two - dimensional autocorrelation function of the surface height @xmath50 : @xmath51 since the equation ( [ eq : gkse ] ) is isotropic , the correlation functions ( [ eq : height - corr ] ) @xmath49 of the resulting surfaces should statistically be independent of the direction of @xmath27 and , thus , depend only on its absolute value @xmath52 .
we therefore define the isotropic height correlation function @xmath53 as averaged over all directions of @xmath27 : @xmath54 fig .
[ fig : surf_smallscale ] shows resulting surface patterns for system size @xmath55 ( or , in lattice units , @xmath56 ) and different values of parameter @xmath3 in ( [ eq : gkse ] ) and the corresponding normalized height correlation functions ( [ eq : height - corr - iso ] ) .
we see that for the surfaces generated by ( [ eq : gkse ] ) with @xmath35 , corresponding to the kse ( [ eq : kse ] ) case , the height correlation function ( [ eq : height - corr - iso ] ) has no maximum ( right panel of fig .
[ fig : surf_smallscale ] ) , just an area of slower decay at distances @xmath28 corresponding to the approximate sizes of cells in the pattern . for @xmath57 ,
the height correlation function obtains a maximum whose distance corresponds to the average size of cells .
we see that this the distance of this maximum increases with @xmath3 , meaning that the cell size increases as the parameter @xmath3 is increased .
( values of @xmath0 coded in gray - scale ) evolving according to ( [ eq : gkse ] ) at system size @xmath56 ( @xmath55 ) with parameters @xmath58 at time @xmath59 ( in the stationary regime ) .
right panel : normalized autocorrelation functions @xmath53 as defined in ( [ eq : height - corr ] ) of the surfaces that are shown on the left panel.,title="fig:",scaledwidth=45.0% ] ( values of @xmath0 coded in gray - scale ) evolving according to ( [ eq : gkse ] ) at system size @xmath56 ( @xmath55 ) with parameters @xmath58 at time @xmath59 ( in the stationary regime ) .
right panel : normalized autocorrelation functions @xmath53 as defined in ( [ eq : height - corr ] ) of the surfaces that are shown on the left panel.,title="fig:",scaledwidth=45.0% ] another thing that can be seen in fig .
[ fig : surf_smallscale ] is that the normalized correlation functions @xmath60 decrease slowly for small values of @xmath61 and faster for larger values .
that is the first indication of the influence of parameter @xmath3 on long - range height correlations .
large - scale height variations in surfaces produced by eq .
( [ eq : gkse ] ) become more distinct as the system size is chosen to be many times larger than the typical cell size .
the examples for some values of @xmath3 can be seen in fig .
[ fig : surf1200 - 01 ] and fig .
[ fig : surf1200 - 02 ] .
coded in gray - scale ) at system size @xmath62 ( @xmath63 ) evolving according to ( [ eq : gkse ] ) with parameters @xmath35 ( left panel ) and @xmath64 ( right panel ) at time @xmath65 ( in the stationary regime).,scaledwidth=50.0% ] coded in gray - scale ) at system size @xmath62 ( @xmath63 ) evolving according to ( [ eq : gkse ] ) with parameters @xmath36 ( left panel ) and @xmath37 ( right panel ) at time @xmath65 ( in the stationary regime).,scaledwidth=50.0% ] by repeating the simulations with different system sizes @xmath8 , one notices a slight dependence of the average saturation value of the surface roughness @xmath32 on the system size .
indeed , the results show that the resulting roughness increases when the system size is increased .
this shows that the resulting surface profile contains spatial fourier components of ever smaller wave number @xmath66 ( and correspondingly larger wavelength @xmath67 ) .
although for small scales ( large wave numbers ) the structures of definite size occur , for larger distances ( small wave numbers ) we get the height variations with long - range dependence , and ( as will be shown in the next two sections ) this dependence has a scale free character .
this section presents some general considerations about the spatial power - spectral densities ( psd ) of surfaces , their connection to the surface roughness @xmath68 , and the effects introduced by the finite system size .
the scaling behavior of @xmath68 when the psd has a power - law shape is derived .
these theoretical results are compared to the numerically calculated scaling properties of @xmath68 for the surfaces generated by ( [ eq : gkse ] ) in the next section .
a two - dimensional surface @xmath50 is a single valued function on the plane @xmath69 . in order to avoid the zero frequency component in the spectrum
, we calculate the spectrum of the surface profile @xmath70 with zero mean : @xmath71 where @xmath72 is the average height of the surface .
fourier transformation @xmath73 of this centered surface profile @xmath70 : @xmath74 here @xmath75 is the wave vector of spatial fourier components of the surface profile . the surface power - spectral density ( psd ) is then defined as : @xmath76 where @xmath8 is the size of the segment of the surface analyzed , i.e. , @xmath77 ^ 2 $ ] .
the integral of the power spectrum @xmath78 over all @xmath79 is equal to the variance of the surface height @xmath80 which by our definition is is equal to the square of the surface roughness @xmath81 : @xmath82 since our model is isotropic , @xmath78 must depend only on the absolute value @xmath83 , and we can get the one - dimensional power spectrum @xmath84 of the surface by integrating the two - dimensional power spectrum over all wave vectors @xmath79 of the same absolute value @xmath66 .
this can be done by expressing the wave vector @xmath79 in the polar coordinates @xmath85 , and then integrating @xmath78 over all angles @xmath86 : @xmath87 from this we get that the one - dimensional psd of the surface can be expressed as : @xmath88 in eq .
( [ eq:1d - psd_def ] ) we have used the fact that for isotropic surfaces @xmath89 .
the integral of this one - dimensional psd @xmath84 over all wave numbers @xmath66 also equals to the square of surface roughness : @xmath90 the surfaces in numerical simulations are represented on a @xmath91 matrix that sets limits to the smallest and the largest possible wave numbers @xmath92 and @xmath93 that can fit into the system , thus , filtering the theoretically defined psd @xmath84 . if @xmath14 is the spatial step size in the simulation ,
then the minimal distinguishable wavelength @xmath94 in the system is approximately equal to this discretization step doubled : @xmath95 and maximal wavelength that can fit into the system is of about double system size : @xmath96 since the distance @xmath28 corresponds to the wave number @xmath97 , we get the minimal and maximal wave numbers for the system : @xmath98 the square of the numerically calculated surface roughness , expressed according to ( [ eq:1d - psd_int ] ) should then be @xmath99 if we keep the discretization step @xmath14 constant , assuming that surface patterns for systems of different sizes ( up to the smallest wave numbers allowed by the system size ) are statistically the same , then , by increasing the system size @xmath16 , according to ( [ eq : minmax_wavenumber ] ) , we reduce the minimal wavenumber @xmath92 in ( [ eq:1d - psd_int - num ] )
. therefore the calculated surface roughness must grow with the system size and its scaling behavior @xmath100 when @xmath16 is increased should be able to give us information about the shape of the surface psd @xmath84 for small wave numbers @xmath101 .
let us assume that the one - dimensional spatial psd @xmath84 ( [ eq:1d - psd_def ] ) of a surface has a power - law dependence on @xmath66 for small wave numbers , smaller than some wavelength @xmath102 : @xmath103 here @xmath104 is some constant , @xmath105 is the shape of the psd spectrum for high wave numbers @xmath66 that does not interest us , since we are interested in large scale behavior of the system and , furthermore , assume that @xmath105 does nt change when the system size is increased .
we also assume that the power for small wave numbers @xmath106
. then , for a system of finite size and @xmath107 , the calculated square of the roughness is expressed by : @xmath108 since @xmath102 and @xmath109 come from the model and @xmath93 comes from numerical scheme , the only variable is @xmath92 which is inversely proportional to the system size .
therefore the second term in ( [ wsquare00 ] ) is just a constant which we denoted by @xmath110 : @xmath111 there are two qualitatively distinct cases .
one is @xmath112 which would result in infinite @xmath81 for a system of infinite size and another case is @xmath113 for which even a system of infinite size would have a finite height variance @xmath81 .
for @xmath112 , the square of the surface roughness @xmath68 from ( [ wsquare00 ] ) grows linearly with the logarithm of the system size @xmath16 : @xmath114 here @xmath104 and @xmath115 with @xmath116 being constant .
the last expression in ( [ wsquare_gamma1 ] ) is more useful , since we do nt know the exact value of @xmath117 .
we see that when @xmath112 , the surface roughness @xmath68 goes to infinity for infinite size system ( @xmath118 ) .
this means that in this case the influence of long - range height variations grows with the system size .
when @xmath113 , from ( [ wsquare00 ] ) we get the following scaling relation : @xmath119 with constants @xmath110 , as defined above , and @xmath120 , which defined as @xmath121 again , since we do nt know the value of @xmath117 , we express the scaling of the surface roughness as @xmath122 with the constants @xmath123 and @xmath124 we see that in this case ( for @xmath113 ) the surface roughness @xmath68 has a finite value for a system of infinite size : @xmath125
in this section we present our numerical results for surfaces generated by ( [ eq : gkse ] ) : the one - dimensional power - spectral densities ( psd ) @xmath84 ( [ eq:1d - psd_def ] ) calculated from the autocorrelation function @xmath53 ( [ eq : height - corr - iso ] ) for system sizes up to @xmath39 and the scaling of the roughness @xmath68 which gives shapes of the @xmath84 , based on the considerations of the previous section
. produced by ( [ eq : gkse ] ) with parameter @xmath35 for system sizes ( in lattice units ) @xmath39 ( blue triangles ) , @xmath56 ( red filled circles ) and @xmath126 ( green squares ) .
the black dashed line represents power - law fit with exponent @xmath127.,scaledwidth=50.0% ] produced by ( [ eq : gkse ] ) with parameter @xmath37 for system sizes ( in lattice units ) @xmath39 ( blue triangles ) , @xmath56 ( red filled circles ) and @xmath126 ( green squares ) .
the black dashed line represents power - law fit with exponent @xmath128.,scaledwidth=50.0% ] since the equation ( [ eq : gkse ] ) is isotropic , the direction of the wave vector @xmath79 does not matter in statistical description of height variations , and in spectral analysis of the surface patterns , the wave number @xmath129 is sufficient to describe the occurring spatial modes .
we can therefore analyze one - dimensional surface spectra @xmath84 , defined in ( [ eq:1d - psd_def ] ) .
we calculate the one - dimensional surface spectrum @xmath84 that depends only on the wave number @xmath66 using the wiener - khinchin theorem which states that the power - spectral density can be obtained from the fourier transform of the autocorrelation function .
although the height correlation function @xmath53 ( [ eq : height - corr - iso ] ) depends only on the absolute value @xmath28 of the shift @xmath27 , it is nevertheless a two - dimensional autocorrelation function of the surface @xmath50 .
thus , by applying the two - dimensional fourier transform on the height correlation function ( [ eq : height - corr - iso ] ) , we get the two - dimensional psd ( [ eq : h_psd ] ) from which we calculate the one - dimensional psd @xmath84 using ( [ eq:1d - psd_def ] ) .
the numerically calculated height correlation function for an isotropic surface @xmath50 has been defined in ( [ eq : height - corr - iso ] ) : @xmath130 here , as before , @xmath131 is the surface height with its average value subtracted , the so - called centered surface profile .
the two - dimensional psd @xmath78 is calculated according to ( [ eq : h_psd ] ) with ( [ eq : h_ft ] ) : @xmath132 by changing the variable @xmath133 and switching the order of integration , we get @xmath134 where , according to ( [ eq : height - corr ] ) , @xmath135 is the height correlation function of the surface @xmath50 .
since the surfaces @xmath50 generated by ( [ eq : gkse ] ) are statistically isotropic , the correlation function ( [ eq : acorr_calc ] ) effectively depends only on @xmath52 and can be denoted by @xmath53 .
using this , expressing the two - dimensional integration in ( [ eq:2d - psd - expl ] ) over @xmath27 in the polar coordinates @xmath136 and integrating over the angular part , we get @xmath137 substituting this into ( [ eq:1d-2d - psd_def ] ) gives the one - dimensional psd of the surface : @xmath138 here @xmath139 is the bessel function of the 1st kind : @xmath140 figures [ fig : specsize0 ] and [ fig : specsize1 ] show the calculated surface spectra of the surfaces generated by ( [ eq : gkse ] ) with parameter values @xmath35 ( the kse case ) and @xmath37 , respectively .
in each figure , the spectra for systems of sizes ( in lattice units ) @xmath126 , @xmath56 and @xmath39 are shown .
each spectrum is obtained by averaging over 6 different realizations .
the figures confirm the above assumption that increasing the system size does not change the small scale structures , since the spectra at large values of @xmath66 coincide .
the appearance of low wave number modes at larger system sizes and the approximate power - law behavior of the spectrum @xmath84 can also be seen . of course , for larger systems , the numerical calculation of the two - dimensional autocorrelation function ( [ eq : height - corr ] ) and surface spectrum ( [ eq : suspec - corr ] ) directly can take a very long time .
therefore , the possibility to obtain @xmath84 from scaling of the surface roughness is very useful . ) with @xmath35 ( filled circles ) and @xmath64 ( filled squares ) .
time averaged square of the surface roughness @xmath81 plotted as a function of the system size @xmath16 ( in lattice units ) .
lines : fits of the results for @xmath35 ( blue line ) and @xmath64 ( red line ) by ( [ wsquare_gamma1 ] ) and ( [ wsquare_gamman1 ] ) respectively , corresponding to the scaling surface spectrum ( [ spectrum_pwl ] ) . resulting fit parameters for @xmath141 :
@xmath142 , @xmath143 ; for @xmath64 : @xmath144 , @xmath145 , @xmath146 . ,
scaledwidth=50.0% ] ) with @xmath36 ( filled triangles ) and @xmath147 ( filled squares).time averaged square of the surface roughness @xmath81 plotted as a function of the system size @xmath16 ( in lattice units ) .
lines : fits of the results for @xmath36 ( blue line ) and @xmath147 ( red line ) by ( [ wsquare_gamman1 ] ) corresponding to the scaling surface spectrum ( [ spectrum_pwl ] ) .
resulting fit parameters for @xmath36 : @xmath148 , @xmath149 , @xmath150 ; for @xmath147 : @xmath151 , @xmath152 , @xmath153 .
, scaledwidth=50.0% ] we have investigated the dependence of this saturated surface roughness on the size of the model system for different values of the parameter @xmath3 in ( [ eq : gkse ] ) .
we investigated how the square of @xmath68 estimated by ( [ w - estimate ] ) depends on the system size @xmath16 ( in lattice units ) for sizes @xmath154 , @xmath155 , @xmath126 , @xmath156 , @xmath56 , @xmath157 , @xmath39 , @xmath158 , @xmath159 , @xmath160 , @xmath18 .
the results are shown in fig .
[ fig : scaling01 ] for @xmath35 and @xmath64 , fig .
[ fig : scaling02 ] for @xmath36 and @xmath37 . for @xmath35 ( the kse case ) , the scaling exponent ( as introduced in ( [ spectrum_pwl ] ) )
@xmath112 fits the results best .
thus the dependence of @xmath81 on @xmath16 can be approximated by ( [ wsquare_gamma1 ] ) and the results should form a straight line in the _ log - linear _ plot , and , indeed they do as can be seen in fig . [
fig : scaling01 ] ( filled black circles ) .
the blue line in fig .
[ fig : scaling01 ] fitted to data for sizes from @xmath155 to @xmath160 . for @xmath64
, the scaling exponent @xmath161 fits the results best and the dependence of @xmath81 on @xmath16 can be approximated by ( [ wsquare_gamman1 ] ) .
[ fig : scaling01 ] shows the calculated results ( filled black rectangles ) .
there seems to be a good agreement between the data and the fitting curve which is fitted to data for sizes from @xmath126 to @xmath160 ( red line in fig .
[ fig : scaling01 ] ) . the results for @xmath36 and @xmath37
give scaling exponents are @xmath162 and @xmath163 , respectively , and the dependence of @xmath81 on @xmath16 can be approximated by ( [ wsquare_gamman1 ] ) .
[ fig : scaling02 ] shows the calculated results ( black triangles for @xmath36 and black rectangles @xmath37 ) together with their respective fits ( blue and red lines in fig .
[ fig : scaling02 ] ) .
the generalized kuramoto - sivashinsky equation ( [ eq : gkse ] ) produces surfaces with disordered cellular patterns ( figs .
[ fig : surf_smallscale ] , [ fig : surf1200 - 01 ] and [ fig : surf1200 - 02 ] ) .
the size of the average size of a cell depends on equation parameter @xmath3 and constitutes a definite scale in the surface pattern ( see the peaks on the correlation functions in [ fig : surf_smallscale ] ) .
however , in larger systems , the long - range height variations of different character become apparent ( figs .
[ fig : surf1200 - 01 ] and [ fig : surf1200 - 02 ] ) .
we investigate these long - range height variations for several values of parameter @xmath3 by calculating the corresponding scaling relations of the surface roughness .
the square of the surface roughness @xmath68 ( [ eq : rough ] ) which is by definition equal to the variance of the surface height @xmath5 , @xmath164 , can also be expressed as an integral ( [ eq:1d - psd_int ] ) of the one - dimensional power spectral density @xmath84 , given in ( [ eq:1d - psd_def ] ) , of the surface over all wave numbers @xmath66 . since the finite size @xmath165 of the system and the discretization step @xmath14 in numerical simulations define the approximate lower and upper cut - off values ( [ eq : minmax_wavenumber ] ) for the possible wave numbers , the estimated value of @xmath81 ( [ eq:1d - psd_int - num ] ) depends on the system size . as indicated by the surface spectra in figs .
[ fig : specsize0 ] and [ fig : specsize1 ] , for systems of increasing size , the small - scale patterns ( corresponding to higher wave numbers @xmath66 ) remain statistically the same , and , additionally , new lower wave number modes arise in larger systems . therefore by increasing the size of the system , from the corresponding change in the calculated value of @xmath81 ( [ eq:1d - psd_int - num ] ) , the shape of the spatial power spectrum @xmath84 can be extracted .
if the behavior of @xmath84 for small wave numbers @xmath66 ( say when @xmath166 , for some @xmath102 ) follows the inverse power - law ( [ spectrum_pwl ] ) , @xmath167 , with @xmath168 , this indicates that long - range , scale - free height variations are present . there can be qualitatively distinct cases for different values of the exponent @xmath109 .
if @xmath169 , then we get the scaling relation ( [ wsquare_gamma1 ] ) which implies that @xmath68 will grow indefinitely as the system size goes to infinity . if , on the other hand , @xmath113 , then we get the scaling relation ( [ wsquare_gamman1 ] )
, therefore @xmath68 will approach finite value ( [ roughinfsiz ] ) as the system size goes to infinity . for the surfaces generated by numerical simulations of the generalized isotropic kuramoto - sivashinsky equation ( [ eq : gkse ] )
, we indeed see the indications that the spatial power spectral density @xmath84 follows a power - law ( [ spectrum_pwl ] ) , since the theoretically calculated scaling relations ( [ wsquare_gamma1 ] ) and ( [ wsquare_gamman1 ] ) fit the numerically established scaling of @xmath81 well ( see figs .
[ fig : scaling01 ] and [ fig : scaling02 ] ) . for the kuramoto - sivashinsky equation ( [ eq : kse ] ) , we get the spectral exponent @xmath112 and the scaling relation ( [ wsquare_gamma1 ] ) . for parameter values
@xmath57 the scaling exponent @xmath109 decreases resulting in the scaling relation ( [ wsquare_gamman1 ] ) giving the surfaces of finite roughness as the size of the system goes to infinity , and thus showing that the generalized kuramoto - sivashinsky equation ( [ eq : gkse ] ) with @xmath57 does not belong to the same universality class as the kse ( [ eq : kse ] ) . | this paper presents numerical results for the two - dimensional isotropic kuramoto - sivashinsky equation ( kse ) with an additional nonlinear term and a single independent parameter .
surfaces generated by this equation exhibit a certain dependence of the average saturated roughness on the system size that indicates power - law shape of the surface spectrum for small wave numbers .
this leads to a conclusion that although cellular surface patterns of definite scale dominate in the range of short distances , there are also scale - free long - range height variations present in the large systems .
the dependence of the spectral exponent on the equation parameter gives some insight into the scaling behavior for large systems . |
hr 8799 is a late a / early f star located at 39.4 pc , with t@xmath1 @xmath0 7350k , m@xmath2 = 1.5 m@xmath3 and a @xmath4 bootis - like pattern of surface metallicity .
the star has received a very high level of attention since the imaging of three massive substellar objects in large radius orbits around it ( marois et al .
hr 8799 and fomalhaut ( kalas et al .
2008 ) are the only two firm examples of systems with separations @xmath5 50 au between the star and planetary - mass companions .
both stars also have luminous planetary debris disks that can provide further insight to the outer structures of their planetary systems ( e.g. , stapelfeldt et al .
2004 ; su et al .
given the bias in current planet - discovery techniques toward close companions , the existence and properties of these widely separated examples open a new perspective on the diversity of planetary systems .
the nature of the companions of hr 8799 depends critically on the age of the system , since their cooling rates ( baraffe et al .
2003 ) enter strongly into the relation of their luminosities to their masses .
the star is near the zero - age main sequence , where traditional methods for estimating its age are not accurate .
marois et al . (
2008 ) used a variety of argumentsm from its debris disk also suggests a young system . ] to estimate an age of a @xmath6 myr and on this basis assign masses of @xmath7 , @xmath8 , and @xmath8 m@xmath9 respectively to hr 8799 b , c , and d. janson et al .
( 2010 ) used vlt / naco to obtain spatially resolved spectroscopy of hr 8799 c from 3.88 to 4.10 @xmath10 m .
although there are some discrepancies in the slope of the spectrum beyond 4 @xmath10 m , it can generally be fitted by a cond model from baraffe et al .
( 2003 ) with t@xmath11 = 1100 k , log g = 4.0 and r@xmath12 = 1.3 r@xmath9 .
these characteristics are consistent with the mass estimate for planet c of 10 m@xmath9 at 60 myr by marois et al .
the minimum mass for deuterium burning , @xmath0 13 m@xmath9 , is often taken as the dividing line between planets and brown dwarfs , in which case these mass estimates indicate that hr 8799 hosts a fascinating planetary system . however , moya et al . ( 2010a ) argue that the stellar age is still unconstrained , pointing out that some of the age estimators are compromised by the @xmath4 bootis character of the star or are statistical in nature .
they used astroseismology to show that an older age of @xmath0 1 gyr is possible , if the inclination of the equatorial plane of the star is @xmath13 . because the masses of the companions are derived from the comparison of the observed luminosity to theoretical evolutionary tracks of giant planets and brown dwarfs , the age of the star is critical in determining these masses . at this older age ,
the companion luminosities would have faded significantly and their assigned masses would be well into the brown dwarf range .
thus , the nature of the hr 8799 system would change significantly .
this would affect the formation scenarios under discussion , and the ability to use these three planets ( of similar age and metallicity ) to calibrate evolutionary and atmospheric models of giant planets . in this letter
, we assess the possibility of more massive companions by considering the limiting case in which the masses of the planets are set to the largest values that dynamical stability can allow ( section [ planet_conf ] ) . given the observed planetary luminosities and using evolutionary models , we estimate the age corresponding to these limiting masses .
we then evaluate the stability of the planets and the dust - producing planetesimals during that timescale to assess whether an older age could still lead to a planetary system ( planets + planetesimals ) consistent with the observations ( section [ stability ] ) .
we discuss the results in section [ discussion ] .
[ intro_planets ]
we now consider whether the configuration of the companions can help resolve the uncertainty over the age of hr 8799 .
the companions have been detected in hst / nicmos archival data taken 10 years prior to the discovery image ( lafreniere et al .
2009 ) , however , their long orbital periods make the determination of the orbital parameters still uncertain . while there are several configurations that would fit ( 1 ) the astrometry , ( 2 ) the observed luminosities , and ( 3 ) the requirement that the system is stable for at least 30 - 160 myr , dynamical analysis by various authors converge on a solution in which the companions are on nearly circular orbits , locked into a double 4:2:1 mean motion resonance ( 1d:2c:4b ) . this possibility was first suggested by fabryck & murray - clay ( 2010 ) , and latter confirmed by godziewski & migaszewski ( 2009 ) and reidemeister et al .
( for a summary of the possible orbital solutions see section 2.4.2 in moro - martn et al .
the companions avoid close encounters in the resonant configuration , ensuring dynamical stability . in this paper , we focus on this configuration because it may be the only solution that can guarantee the dynamical stability of the system on a timescale comparable to the stellar age , and in the case of massive planets .
the analysis in fabrycky & murray - clay ( 2010 ) concluded that , under this configuration , the planetary masses can be increased up to 1.9 times their nominal values ( see their figure 11 ) .
therefore , we consider the following planetary masses with orbital elements listed in table [ planetparam ] : 1.27@xmath1410@xmath15 m@xmath16 (= 13.30 m@xmath9 ) for planet b , and 1.80@xmath1410@xmath15 m@xmath16 (= 18.85 m@xmath9 ) for planets c and d ; this would put planets c and d well into the brown dwarf regime , while planet b would be near the transition .
following the labeling convection in moro - martn et al .
( 2010 ) , we will refer to this configuration as fit d6 .
lccccccc 1.27@xmath1410@xmath15 ( b ) & 67.91 & 0.002 & 0 & 0 & @xmath17 & @xmath17 + 1.80@xmath1410@xmath15 ( c ) & 37.97 & 0.005 & 0 & 0 & 0 & 0 + 1.80@xmath1410@xmath15 ( d ) & 23.52 & 0.083 & 0 & 0 & @xmath17 & 0 + the conclusion that the values adopted here are the maximum permitted masses is based on the analysis by beauge et al .
( 2003 ) : for a given pair of eccentricities ( @xmath18 ) , they estimated the maximum mass of the innermost planet ( @xmath19 ) for which stable apsidal corotation at the 2:1 resonance can exist . for the configuration considered in this paper ( see table [ planetparam ] ) , where the planet pair b - c has @xmath20 = 0.005 and @xmath21 = 0.002 , beauge et al .
( 2003 ) estimated that the maximum mass of @xmath22 is @xmath23 0.012 m@xmath2 ( see their figure 8) ; similarly , the planet pair c - d , with @xmath24 = 0.083 and @xmath21 = 0.005 , would correspond to @xmath25 having a maximum mass of @xmath0 0.013 m@xmath2 .
these values are similar to those adopted in this paper for planets c and d ( 1.80 @xmath1410@xmath15 m@xmath16 = 0.012 m@xmath2 , for m@xmath2= 1.5 m@xmath16 ) .
configurations with planetary masses more massive than these can not have a stable corotation orbit in the 2:1 resonance .
in other words , the fit d6 adopted in this study is a limiting case : the upper bound for the dynamically allowed planetary masses .
using different evolutionary models , and based on the luminosities reported by marois et al .
( 2008 ) , log(@xmath26 ) = -5.1 for planet b and log(@xmath26 ) = -4.7 for planets c and d , we can estimate the age of the system that would correspond to the observed luminosities and the masses adopted in this paper : using the models of chabrier et al .
( 2000 ) , the age would be 100500 myr ( it can not be constrained any further because their tables do not include results for intermediate ages ) ; from baraffe et al .
( 2003 their figure 1 ) , we obtain an age of 250350 myr ; while from burrows et al .
( 1997 - their figure 7 ) we get that if planet b were to have 1.27@xmath1410@xmath15 m@xmath16 , its observed luminosity would correspond to a stellar age of @xmath0 260 myr .
the temperature of 1100 k for hr 8799 c ( janson et al .
2010 ) can also be used to constrain the age : from chabrier et al .
( 2000 - their figure 2 ) , this temperature and adopted mass would correspond to an age of @xmath23 225 myr , while from baraffe et al .
( 2003 - their figure 1 ) the age would be in the range 250280 myr .
we conclude that if the planets were to have the limiting masses adopted in this paper ( 1.9 times their nominal values ) , their current luminosities would imply a system age of approximately 250350 myr .
we now evaluate the stability of the planets and the dust - producing planetesimals on the timescales estimated above .
following fabrycky & murray - clay ( 2010 ) , and to assess how sensitive the results are to small changes in initial conditions , we consider several planetary configurations around the one described in section [ planet_conf ] and table [ planetparam ] ; the differences are 10@xmath27 au for the semi - major axes , 10@xmath27 for the eccentricities , and 0.01@xmath28 for the inclinations , ascending nodes , longitudes of periastron , and mean anomalies . in total , we model 14 different sets of initial conditions for the orbits .
in addition to the three companions and the central mass of 1.5 m@xmath16 , the simulations include 500 test particles .
seven of the simulations consider particles in the inner disk , uniformly distributed between 2 au and the semi - major axis of the inner - most planet , while the remaining seven consider particles in the outer disk , spaced uniformly between the semi - major axis of the outermost planet and 300 au .
we assume that the planets and the dust - producing planetesimals formed out of a thin disk and are co - planar .
the particles are on initially circular orbits , with angular elements chosen randomly between 0 and 2@xmath17 .
particles are removed if they approach the star closer than 2 au , or approach a planet closer than its hill radius .
the orbits were integrated using the multiple time step symplectic method skeel - symba ( duncan , levison & lee 1998 ) . .
from top to bottom , the panels correspond to semi - major axis ( @xmath29 ) , eccentricity ( @xmath30 ) and resonant angles for the 2:1 resonance between planets c and d ( @xmath31 ) , and the 2:1 resonance between planets b and c ( @xmath32 ) .
the mean and standard deviation of the resonant angles are listed in table [ lifetime ] .
the colors correspond to planet b ( black ) , c ( orange ) and d ( blue ) .
_ right _ : results from the dynamical simulation of 500 test particles in the inner planetesimal disk .
top : test particle s lifetimes .
middle : allowed parameter space for the planetesimals orbital elements , where the shaded areas indicate regions where test particle s orbits are unstable due to planet - crossing ( _ striped area _ ) or overlapping first order mean motion resonances ( _ dotted area _ ) ; the squared symbols show the maximum eccentricity attained by test particles on initially circular orbits .
bottom : number of surviving test particles at the onset of dynamical instability . ]
but for a simulation of 500 test particles in the outer planetesimal disk . in this case ,
the high baseline value of the test particles eccentricities and slight rising trend is a numerical artifact from using stellar - centric osculating elements .
] figures [ ploe1 ] and [ ploe2 ] show the evolution of the planets semi - major axes , eccentricities , and resonant angles for the 2:1 resonance between planets c and d ( @xmath31 ) , and the 2:1 resonance between planets b and c ( @xmath32 ) , for the two longest - lived simulations , one for the inner disk and one for the outer disk , with dynamical lifetimes of 137.5 myr and 147.5 myr , respectively .
while the companions are locked into the double resonance , the system is stable .
table [ lifetime ] lists the resonant angles ( mean and standard deviation ) and lifetimes of the 14 simulations ran , the latter ranging from of 9 to 147 myr . using the mercury integrator ,
fabrycky & murray - clay ( 2010 ) found that the time to instability of their six simulations ( with the same set of initial conditions ) range from 25.9 myr to 102.6 myr .
we conclude that if the planets were to have the limiting masses adopted in this paper ( 1.9 times their nominal values ) , the lifetime of the system would be @xmath23 150 myr .
following moro - martn et al .
( 2007 , 2010 ) , we have estimated the potential location of the dust - producing planetesimal belts under the configurations considered in this paper .
the possible niches are the regions where the orbits are stable and the maximum eccentricities excited by the planetary perturbations remain low ( @xmath33 0.3 ) , ensuring the survival of planetesimals against quick erosion from mutual collisions .
figures [ ploe1 ] and [ ploe2 ] show the results for the two longest - lived simulations already discussed ( one for the inner disk , and one for the outer disk ) .
we find that the outer edge of the inner ( warm ) planetesimal disk is r@xmath34 9 au , and that the inner edge of the outer ( cold ) planetesimal disk is r@xmath35 150 au .
we have adjusted the disk model presented in su et al .
( 2009 ) to this new geometry and find that a satisfactory fit to the spectral energy distribution can be achieved readily .
( for comparison , the configuration adopting the nominal masses of 7 , 10 and 10 m@xmath9 for planets b , c and d , respectively
labeled d1 in moro - martn et al .
2010 results in r@xmath34 12 au and r@xmath36 110 au ) .
we conclude that the presence of planets with masses 1.9 times their nominal values would still be consistent with the sed and imagining of the debris disk around hr 8799 .
lcccccccccccccc dyn .
lifetime ( myr ) & 137.5 & 41.2 & 110.9 & 59.3 & 29.9 & 86.5 & 32.0 & 117.3 & 147.5 & 43.3 & 19.2 & 9.1 & 17.6 & 31.8 + mean(@xmath37 ) ( @xmath28 ) & -0.04 & -0.84 & 0.14 & -0.50 & 6.52 & -0.06 & -4.82 & 0.86 & 0.19 & -1.64 & 1.03 & 0.29 & 0.67 & 0.56 + @xmath38 ) ( @xmath28 ) & 44.4 & 45.7 & 44.6 & 43.9 & 48.7 & 44.1 & 48.9 & 44.3 & 43.4 & 46.4 & 103.2 & 103.5 & 103.1 & 102.5 + mean(@xmath31 ) ( @xmath28 ) & -0.08 & -0.17 & -0.20 & -0.08 & 1.90 & 0.03 & -0.96 & 0.24 & 0.01 & -0.13 & -5.67 & 3.47 & 0.80 & -0.56 + @xmath39 ( @xmath28 ) & 8.9 & 9.9 & 9.2 & 9.4 & 10.3 & 9.0 & 9.9 & 9.2 & 8.6 & 10.1 & 103.0 & 104.2 & 101.0 & 102.0 +
the configuration considered in this paper ( d6 ) is a limiting case , in the sense that the adopted companion masses are the maximum values for which the stabilization mechanism , the double 4:2:1 resonance , can apply .
based on stability considerations , it is therefore unlikely that the companions are more massive than considered here . in moro - martn
( 2010 ) we found that configurations in which the planets adopt their nominal masses ( 7 , 10 and 10 m@xmath9 for planets b , c and d , respectively labeled d1 , d4 and d5 )
are consistent with planet and dust observations because : ( 1 ) the planets masses correspond to an age estimate of @xmath0 60 myr and the system is found to be stable for @xmath5 160 myr ; and ( 2 ) the orbits of the dust - producing planetesimals fulfill two requirements : ( 2a ) they are in agreement with the dust location as derived from debris disk observations ; and ( 2b ) for the timescale during which the planet orbits are stable , the planetesimal orbits are also stable and their eccentricities remain low ( implying survival against erosion ) . in contrast , the configuration considered in this paper , where the planets are 1.9 times more massive , is not fully consistent . conditions ( 2a ) and ( 2b ) above are fulfilled : for the timescale during which the planets are on stable orbits , there are planetesimals in the regions where the dust is inferred to be located , and the planetesimal orbits are stable and their eccentricities remain low . however , condition ( 1 ) is violated : while the substellar object evolutionary models predict an age of the system around 250350 myr ( baraffe et al .
2003 ) or @xmath0 260 myr ( burrows et al .
1997 ) from luminosity constraints , or @xmath23 225 myr ( chabrier et al . 2000 ) and 250280 myr ( baraffe et al . 2003 ) from temperature constrains , the dynamical lifetime of the system is significantly smaller , @xmath23 150 myr . based on dynamical considerations
we therefore conclude that the large companion masses considered in this paper are unlikely and that the dynamical state of the system favors ages of @xmath0 100 myr and companion masses slightly below the brown dwarf regime .
this conclusion is supported by the strong emission of the debris disk at 24 @xmath10 m , behavior that is very uncommon around stars older than @xmath0 500 myr ( gaspar et al .
in addition , the average temperature of 7350 k and the luminosity of 5 l@xmath3 from sadakane ( 2006 ) and moya et al . ( 2010b ) places the star at or below the zero age main sequence and at a very young age compared with other @xmath4 boo stars ( paunzen et al .
. however , these results are not consistent with the possibility that the system is @xmath0 1 gyr old .
the astroseismology modeling that suggests this age ( moya et al .
2010a ) depends critically on the rotation speed of the star and hence on the assigned inclination .
it is only valid if the inclination of the equatorial plane of the star is @xmath4036@xmath28 .
the case for @xmath41 results in two sets of possible solutions , a stellar mass of 1.321.33 m@xmath16 and an age of 11231625 myr , or a stellar mass of 1.441.45 m@xmath16 and and age of 26430 myr ( with 16.7% of the models lying in the generally adopted age range of 30160 myr ) . for @xmath42 , moya et al .
( 2010a ) obtain a stellar mass of 1.32 m@xmath16 and and age of 11261486 myr .
however , lafreniere et al .
( 2009 ) suggested that the inclination of the orbital plane of the planets is @xmath43 = 1323@xmath28 from astrometry observations with a 10 year baseline ( using nicmos archival data ) .
( 2009 ) reported from the degree of symmetry of the spatially resolved 70 @xmath10 m disk that the inclination of the disk is consistent with being face - on and unlikely to be @xmath44 25@xmath28 .
we have re - evaluated these observations and place a 3@xmath45 upper limit to @xmath46 of 40@xmath28 .
upcoming jcmt / scuba-2 , herschel and hst imaging of hr 8799 will be able to further constrain the inclination of the debris disk .
if we assume that the equatorial plane of the star coincides with the orbital plane of the planets and the dust - producing planetesimals ( i.e. @xmath47 ) , these observational constraints indicate that indeed the inclination of the star is likely inside the 18 @xmath48 range where the astroseismology models can not be applied , or that it is close to 36@xmath28 where they permit solutions consistent with the @xmath0 100 myr age derived from other considerations .
we thank hal levison for providing skeel - symba for the dynamical simulations and m. janson , d. fabrycky , r. murray - clay , a. moya and p. kalas for useful discussion . a.m.m .
acknowledges funding from the spanish micinn ( ramn y cajal program ryc-2007 - 00612 , and grants aya2009 - 07304 and consolider ingenio 2010csd2009 - 00038 ) .
baraffe , i. , chabrier , g. , barman , t. s. , allard , f. , & hauschildt , p. h. 2003 , , 402 , 701 beaug , c. , ferraz - mello , s. , & michtchenko , t. a. 2003 , , 593 , 1124 burrows , a. , et al .
1997 , , 491 , 856 chabrier , g. , baraffe , i. , allard , f. , & hauschildt , p. 2000 , , 542 , 464 duncan , m. j. , levison , h. f. , & lee , m. h. 1998 , aj , 116 , 2067 fabrycky , d. c. , & murray - clay , r. a. 2010 , apj , 710 , 1408 gspr , a. , rieke , g. h. , su , k. y. l. , balog , z. , trilling , d. , muzzerole , j. , apai , d. , & kelly , b. c. 2009 , , 697 , 1578 godziewski , k. , & migaszewski , c. 2009 , mnras , 397 , l16 janson , m. , bergfors , c. , goto , m. , brandner , w. , & lafrenire , d. 2010 , , 710 , l35 lafrenire , d. , marois , c. , doyon , r. , & barman , t. 2009 , apjl , 694 , l148 marois , c. et al .
2008 , science , 322 , 1348 moro - martn , a. , malhotra , r. , bryden , g. , rieke , g. h. , su , k. y. l. , beichman , c. a. , & lawler , s. m. 2010 , , 717 , 1123 moro - martn , a. , et al .
2007 , , 668 , 1165 moya , a. , amado , p. j. , barrado , d. , garca hernndez , a. , aberasturi , m. , montesinos , b. , & aceituno , f. 2010a , , 405 , l81 moya , a. , amado , p. j. , barrado , d. , hernndez , a. garca , aberasturi , m. , montesinos , b. , & aceituno , f. 2010b , , 406 , 566 paunzen , e. , handler , g. , weiss , w. w. , nesvacil , n. , hempel , a. , romero- colmenero , e. , vuthela , f. f. , reegen , p. , shobbrook , r. r. , & kilkenny , d. 2002 , a&a , 392 , 515 reidemeister , m. , krivov , a. v. , schmidt , t. o. b. , fiedler , s. , mller , s. , lhne , t. , & neuhuser , r. 2009 , , 503 , 247 sadakane , k. 2006 , pasj , 58 , 1023 stapelfeldt , k. r. , et al .
2004 , apjs , 154 , 458 su , k. y. l. , et al . 2009 ,
apj , 705 , 314 | we consider the limiting case for orbital stability of the companions to hr 8799 .
this case is only consistent with ages for the system of @xmath0 100 myr , not with the 1 gyr age proposed from astroseismology .
the discrepancy probably arises because the inclination of the star is smaller than assumed in analyzing the astroseismology data .
given this young age , the best estimates of the companion masses place them by a small margin on the planet side of the division between planets and brown dwarfs . |
detecting dust emission associated with the diffuse ionized phase is a complex task due to two main competing factors .
first of all , large grains are expected to be selectively destroyed in the warm ionized medium ( wim ) due to the passage of multiple shock fronts which would reduce their emission .
in addition , along most line of sights ( los ) , the diffuse ionized gas is likely mixed with the other phases of the gas ( atomic and molecular ) , and its contribution is not expected to be dominant at any latitude . because of these difficulties
, contradictory results concerning the presence of dust in the wim have been reported in the literature .
@xcite , looking at the high galactic latitude residuals of the ir / hi correlation between 100 and 1000 @xmath0 , claim a statistical detection .
in particular , by adopting an emissivity spectral index of 2 , they find an emissivity law in the atomic and ionized gas phase of , respectively , @xmath6 @xmath7 and @xmath8 @xmath7 , with t@xmath9 = 17.5 k and t@xmath10 = 20 k. following this preliminary investigation , @xcite extend their analysis , and carry out a decomposition of the far infrared ( fir ) emission using wham @xcite h@xmath2 data as a tracer of the ionized medium .
the result of this study confirms the early detection , with an emissivity law for the ionized phase of @xmath11 @xmath7 , for a fixed temperature of 17.2 k for both the atomic and ionized phase . at low galactic latitudes , @xcite derive dust emissivities associated with the atomic , molecular and ionized gas phases , using an inversion method applied to iras ( neugebauer et al .
1984 ) , dirbe ( hauser 1993 ) , archeops ( benoit et al . 2002 ) and wmap ( bennett et al . 2003 ) data .
according to their analysis , the average dust temperatures in the atomic and molecular phases are comparable , and of order of 19.8 and 19.2 k respectively , while dust in the ionized phase is significantly warmer , t@xmath10 @xmath12 26.7 k. at the same time , @xcite looking at the dirbe / wham correlation in five selected high - latitude galactic regions find , contrary to @xcite , that the h@xmath2-correlated dust component is negligible in all cases . despite the observational caveats , assessing
the presence of dust in the ionized gas phase is very important , given that the optical properties of dust grains can vary with the properties of the local environment . to date , no clear - cut indication of such a behaviour has been found though .
@xcite perform a statistical study of all the molecular clouds in the lmc and do not find any systematic change in dust abundances nor in dust temperature between the molecular and the atomic phase . on the contrary , @xcite and @xcite , by looking at nearby galactic clouds ,
do report evidence of this kind of variations . in the ionized medium , due to the proximity of the ionized gas with uv sources such as o and b stars and the cumulative effect of shock fronts , changes in the optical dust properties
are indeed expected .
for instance , depletion of pahs has been observed toward individual hii regions @xcite .
however , this phenomenon does not appear to be systematic , and its origin is still poorly understood . in this paper , we present a systematic investigation of the properties of dust associated with the ionized gas in the lmc .
the lmc , located at a distance of @xmath3 50 kpc @xcite , has an advantageous almost face - on viewing angle . at -32@xmath13 latitude
, the contamination along the los is expected to be small compared to the plane of the galaxy . for this galaxy ,
@xcite carried out a cross - correlation study of the combined spitzer sage @xcite and iras data , with hi and co data .
that work allowed to characterize ( in terms of emissivities , abundances , etc . ) the properties of dust mixed with the atomic and molecular medium , as well as to point out the existence of an excess of emission at 70 @xmath0 .
interestingly , @xcite showed that this excess appears to be spatially correlated with both the atomic and the diffuse ( i@xmath14 200 r ) ionized gas .
the present study builds over and expands the analysis described in @xcite , by including the ionized medium in the cross - correlation and by encompassing different physical regimes of this phase of the gas , from the most tenuous environments ( diffuse medium ) , to the densest ones ( hii regions ) .
the paper is organized as follows .
section [ sec_data ] summarizes the data we used . in section [ sec_decomp ]
, we describe the decomposition method as well as the dust emission modeling .
we present our main results in section [ sec_res ] .
concluding remarks are provided in section [ sec_cl ] .
along the lines of @xcite , we use , as a tracer of dust emission from the nir to the fir , the spitzer sage data .
the data span the wavelength range 3.6 @xmath0 to 160 @xmath0 , with an angular resolution from @xmath15 to 40@xmath16 , respectively .
these spitzer data are combined with the 12 and 100 @xmath0 iris data ( improved reprocessing of the iras survey , * ? ? ?
* ) characterized by an angular resolution close to 4@xmath17 .
the inclusion of the irac 8@xmath0 and iris 12 @xmath0 data allows us to take into account the contribution from the most prominent pah features , i.e. at 7.7 and 11.3 @xmath0 , while with mips 24 and 70 @xmath0 the bulk of the emission from vsgs is incorporated .
in addition , the 100 @xmath0 iris data point , combined with the mips 160 @xmath0 measurement , provides the necessary leverage for the derivation of the equilibrium temperature of the bgs . as a tracer of atomic gas , we use the @xcite 21-cm map , which is a combination of interferometric data obtained with the australia telescope compact array ( atca ) , at 1@xmath17 angular resolution , and the parkes antenna , at 15.3@xmath17 angular resolution @xcite . assuming the gas is optically thin , the column density can be derived using the relation : @xmath18 where @xmath19 is the hi integrated intensity map and the conversion factor @xmath20 is taken equal to @xmath21 @xmath22 @xcite .
the integrated map was built by summing up all the velocity channels from 190 to 386 km s@xmath23 .
for the purpose of estimating and subtracting the galactic foreground contribution to each ir wavelength , we use a galactic hi column density map constructed by @xcite .
more details on this procedure can be found in @xcite .
the @xmath24co ( j=1 - 0 ) survey obtained with the 4-m radio nanten telescope of nagoya university at las campanas observatory , chile ( see * ? ? ?
* ) is adopted as a tracer of the molecular gas . in particular , we make use of the second nanten survey which , characterized by an angular resolution of 2.6@xmath17 , evidenced the existence in that galaxy of 272 individual molecular clouds @xcite . we compute the molecular column densities using the relation : @xmath25 where @xmath26 is the co integrated intensity map , generated by integrating the original co cube over the 200 @xmath27 v@xmath28 320 km s@xmath23 velocity range .
@xcite derived the @xmath29 conversion factor assuming that each cloud is at virial equilibrium .
they found an average value for all the clouds of @xmath30 h@xmath31/@xmath32/(k km s@xmath23 ) . from cloud
to cloud , values vary from 1.6@xmath33 to 4.2@xmath34 @xmath35 . their average value is larger than that estimated by @xcite of @xmath36 @xmath35 , using the mopra telescope data , at a resolution of 33@xmath37 .
this difference is due to the fact that low co brightness clouds have been excluded from the mopra survey , and the @xmath38 value has been deduced from a smaller number of clouds compared to nanten observations . in this work , in order to be consistent with our previous studies ( e.g. * ? ? ?
* ) , we use the @xmath29 values determined for each individual cloud by @xcite . for clouds for which the virial mass could not be derived , we adopt the lmc average value of @xmath29=@xmath39 h@xmath31/@xmath32/(k km s@xmath23 ) . for a comprehensive tabulation of @xmath29 values ,
we refer the reader to @xcite ( see their table 5 ) .
the warm ionized gas , characterized by an electron temperature of the order of @xmath41 k , emits both recombination lines and radio free - free continuum . among the observed recombination lines ,
the h@xmath2 line is the brightest and is due to a transition of the balmer line series at 6562.81 @xmath42 . as for free - free emission ,
this consists of free electrons being scattered and decelerated by their encounters with ions .
h@xmath2 and radio templates are available for the lmc and both tracers are used in our analysis to assure an independent cross - check of the results . for the h@xmath2
, we use the southern h - alpha sky survey atlas ( shassa , * ? ? ?
* ) data set which covers the southern hemisphere ( @xmath43 ) , at an angular resolution of @xmath44 , and with a sensitivity of 2 rayleigh / pixel ( 1 r=10@xmath45 photons @xmath7 s@xmath23 ) .
following @xcite , if we assume that the electron density @xmath46 is constant along each los , the @xmath47 column density can be derived using the relation : @xmath48 where 1 r=2.25 pc @xmath4 for t@xmath49=8000 k ( see , for instance * ? ? ?
* ) . for @xmath46
, we adopt different values for different regimes of the ionized gas , as described in section [ sec_reg ] .
to trace the radio free - free emission , we use observations obtained with the parkes 64-m telescope at 4.75 ghz ( see * ? ? ?
* for a complete description of the parkes survey ) . at that frequency
the beam size is 4.9@xmath17 , but after correcting for scanning effects , the effective beam size results to be equal to @xmath50 . in order to compare the radio and h@xmath2 maps and compute gas column densities for the radio map
, we transform the radio emission to h@xmath2 units , using equation 11 in @xcite : @xmath51 where @xmath52 ( brightness temperature ) is in @xmath53 and @xmath54 in rayleigh .
@xmath55 is @xmath56 in units of @xmath57 k , while the factor @xmath58 depends on the frequency and the electron temperature ( see table 3 in * ? ? ?
* ) and @xmath59 is the frequency in ghz . for @xmath60 8000 k ( @xmath61 ) , @xmath58 is close to unity . to apply equation [ eq_tb ] , we first convert the original units of the parkes map ( mjy / beam ) into jy , and these into t@xmath62 using the expression : @xmath63 with @xmath64 the total flux density at 4.75 ghz , c the speed of light , @xmath65 the frequency and k the boltzmann constant .
once the parkes map has been converted into @xmath66 units , the h@xmath67 column density can be computed by applying equation [ eq_halpha ] .
the sensitivity of the parkes telescope is 8 mjy / beam , also equivalent to 28 r , i.e. significantly lower than the sensitivity threshold of the shassa map .
this implies , as we will also see in section 3.2 , that while the shassa map will be used as a tracer of the ionized gas for all types of environmental conditions ( i.e. from the diffuse to the dense ionized medium ) , the parkes map will only be used to probe the densest / brightest regions . for the pixels corresponding to bright hii regions ,
we have performed a comparison of the radio parkes map , converted into h@xmath2 emission ( @xmath68 ) , with the shassa map .
the comparison highlights that , for 88@xmath69 of the pixels , the h@xmath2-parkes data are higher than the shassa one .
the median value of the ratio between the two data sets is 1.6 .
to explain such a discrepancy , we invoke extinction as a source of attenuation in the shassa map . in this respect , we remind the reader that h@xmath2 measurements are affected by dust absorption , while radio data provide an uncontaminated observational window on the ionized gas . in the lmc
, extinction is usually considered rather low .
for instance , @xcite and @xcite quote an average a(v ) of 0.3 mag . in our analysis ,
the factor 1.6 of discrepancy would induce an extinction value in the shassa map of 0.51 mag in bright hii regions .
following @xcite , @xcite estimate the absorption in h@xmath2 as @xmath70 .
this result is obtained using an extinction curve for the milky way . however , as shown by @xcite , the extinction curve in the lmc is very similar to the galactic one .
therefore , the estimate above also applies here . in that case
, the corresponding extinction of lmc bright hii regions in the visible is a(v)=0.63 mag .
this value is slightly higher than the extinction estimated using the lmc a(v ) map constructed by @xcite , and weighted by the h@xmath2 emission ( @xmath71 mag ) over the same hii regions .
it is unlikely that bright hii regions are all confined behind dust clouds , which may indicate that the extinction could have been underestimated when constructing the lmc a(v ) map from star counts , as already suggested in @xcite . in the following ,
we therefore expect differences of about a factor 1.6 when comparing results using shassa and parkes data ( see section [ sec_case3_results ] ) .
all the data used for the analysis ( ir , hi , co , h@xmath2 ) have been degraded to the lowest angular resolution of the parkes 4.75-ghz map , using a gaussian kernel with a fwhm of 5.6@xmath17 .
after being convolved , the maps have been reprojected onto the same grid , with a 2@xmath17 pixel size .
finally , point source subtraction has been applied to all the ir maps ( see * ? ? ? * for the description of the map processing ) .
we consider three regimes of the ionized gas , i.e. from diffuse h@xmath72 to bright hii regions , and for each regime we investigate the dust properties independently , due to potentially intrinsic variations in dust composition , evolution and physics of the grains .
to define each regime , we set three thresholds in h@xmath2 emission intensity using the shassa map , convolved to 5.6@xmath17 resolution , and the catalog of hii regions in the lmc compiled by @xcite ( hereafter kh86 ) .
we start by analyzing the pixel intensity distribution in the shassa map ( see figure [ fig_ne ] ) .
the mean ( @xmath73 ) of the distribution is 1.4 r , with a dispersion ( @xmath74 ) of 1.1 r and a minimum of 0.4 r , which is the sensitivity of the shassa data for the 5.6@xmath17 angular resolution and the 2@xmath17 pixel size of the map .
about 99@xmath69 of the pixels lies in the range 0.4 r @xmath75 6.7 r , corresponding to @xmath76 5@xmath74 .
we notice that none of the sources in the kh86 catalog are located in regions of the shassa map where the h@xmath2 intensity is within this range ( see figure [ fig_cases ] ) , and therefore , we adopt it to define the _ diffuse _ ionized gas regime . for @xmath77 6.7 r , we look , in the shassa map , at the intensity distribution of the pixels which correspond to the location of the hii regions in the kh86 list . to select the pixels , we use , for each cataloged source , its coordinates and measured angular diameter . the h@xmath2 intensity distribution obtained by this process
is shown in figure [ fig_ne ] ( insert panel ) .
this is characterized by a mean value of 26.3 r , with a dispersion of 17.4 r. in this case , 99@xmath69 of the pixels lies in the range 6.7 r @xmath75 113.3 r , and such a range is used to define what , in this context , we denote as _ typical _ hii regions . for @xmath78 113.3 r
, we have a few very bright hii regions which are listed in the kh86 catalog .
we set an upper limit of 1000 r to exclude extremely bright sources , such as 30 doradus , which due to their complexity - would require a dedicated analysis .
the range 113.3 r @xmath75 1000 r hence identifies the third regime . in summary , we have three distinct regimes / cases which we will be tackled by our analysis : * case 1 : 0.4 r @xmath75 6.7 r ( diffuse ionized gas ) ; * case 2 : 6.7 r @xmath75 113.3 r ( typical hii regions ) ; * case 3 : 113.3 r @xmath75 1000 r ( very bright hii regions ) . for each regime
, we need to adopt a characteristic @xmath79 in order to compute the h@xmath80 column density from equation [ eq_halpha ] . for the diffuse component ,
no estimate is available from the literature .
we have made an attempt to derive this value directly from the shassa map .
for this purpose , we first compute , for each pixel in the range 0.4 r @xmath75 6.7 r , the emission measure ( em ) , by applying equation [ eq_tb ] in @xcite and assuming t@xmath81 = 8000 k. to derive the rms electron density from the average em so computed ( @xmath82 = 5.3 @xmath83 0.1 pc @xmath4 ) , we need to assume a linear scale of the emitting ionized medium .
our 5.6@xmath17-shassa map has a pixel size of 2@xmath17 . at a distance of 50 kpc
, this corresponds to a linear size of @xmath3 30 pc .
however , such a value is very likely a severe underestimate of the actual vertical extent of the diffuse ionized gas .
in fact , kh86 have analyzed the distribution in linear diameters for their cataloged hii regions , and obtained that the largest sources in the sample have a size of about 400 pc . in the hypothesis that the diffuse gas extends far beyond the layer of discrete hii regions ,
if we denote with @xmath84 the vertical extent of h@xmath80 and take @xmath84 = 500 pc , we obtain an average electron density n@xmath85 of 0.10 @xmath83 0.08 @xmath86 .
this value is close to what has been derived for our galaxy , i.e. n@xmath81 @xmath3 0.03 0.08 @xmath86 , for an estimated scale height ranging between 1000 and 1800 pc @xcite . if , instead of 500 pc , we assume @xmath84 = 1000 pc , we retrieve n@xmath85 = 0.07 @xmath83 0.06 @xmath86 , matching the galactic value .
despite such a result , given the uncertainty in @xmath84 for the lmc , we choose to adopt the average value derived from galactic measurements .
therefore , for case 1 , we take n@xmath85 = 0.055 @xmath86 . for cases 2 and 3 ,
we compute em , for each hii region in the kh86 catalog by applying equation 4 in @xcite . for each source , we use the quoted h@xmath2 flux and we set t@xmath81 = 8000 k. the measured h@xmath2 flux is converted into rayleighs , and the intensity cuts at 113.3 r and 1000 r for , respectively , case 2 and 3 are applied .
the averaged emission measures for the two cases ( @xmath87 = 183.8 @xmath83 17.8 pc @xmath4 , @xmath88 = 1322.4 @xmath83 167.7 pc @xmath4 ) are then used to derive an rms electron density . following this procedure ,
we obtain : n@xmath89 = 1.52 @xmath83 0.44 @xmath86 and n@xmath90 = 3.98 @xmath83 2.05 @xmath86 , i.e. 20 and 80 times larger than the diffuse component . [ cols="<,^,^,^,^,^,^ " , ] in the lmc , @xcite evidenced the existence of a large fir excess with respect to the atomic , molecular and ionized gas phases , which they attributed to either molecular gas with no associated co emission or to optically thick hi .
they showed that this component is spatially correlated with hi .
@xcite performed a statistical analysis of lmc molecular clouds and showed that if the fir excess was not taken into account in the ir correlations , the dust abundance in the atomic phase would be systematically overestimated by a factor 1.7 ( see their section 3 ) .
following @xcite , we decompose the ir emission at each wavelength and for each regime , as a function of the atomic , molecular and ionized gas column density : @xmath91 with @xmath92 the column density in the fir excess component . in the following , we refer to the sum of the atomic and fir excess components as the atomic phase . despite this working hypothesis
, it is important to keep in mind that the fir excess could be molecular in nature , instead of atomic , and associated with warm h@xmath31 gas @xcite .
the parameters a , b , c , denoting the emissivities associated with each phase of the gas at a given wavelength , and the constant d are determined using regression fits . to remove any possible offsets in the data ,
we subtract a background in all the maps .
the background is computed as the median over a common area , corresponding to the faintest 10@xmath69 of the hi data .
this step also ensures to have a null ir emission for a null column density .
the h@xmath2 data are used for all regimes of ionized gas described in the previous section , while , due to the limited sensitivity of the 4.75-ghz map , we can only use the radio data for case 3 . as a consequence
, a direct cross - check of the results obtained by adopting both tracers will only be possible for this case .
although case 3 corresponds to bright hii regions for which , presumably , most of the radio emission is due to thermal free - free , yet we might have a residual non - thermal contamination associated with synchrotron radiation . to investigate the synchrotron contamination ,
we complement the 4.75-ghz parkes data with the 1.4-ghz data , obtained with the same telescope .
both maps are smoothed to the common resolution of 16.6@xmath17 , characterizing the 1.4-ghz data .
following @xcite , we estimate the fraction of emission contributed by free - free as : @xmath93 where @xmath94 is the thermal fraction at 4.75 ghz , @xmath95 and @xmath96 are the total flux densities at 4.75 and 1.4 ghz , and @xmath97 and @xmath98 are the free - free and synchrotron spectral indices . following @xcite and @xcite , we take @xmath97 = 0.1 . the spectral index of synchrotron radiation
is known to present large spatial variations . in our galaxy , it ranges between 0.6 in star - forming complexes to 1.2 in the interarm regions .
if , in equation [ eq_sync ] , we adopt a fixed value of @xmath98=0.6 , we derive an average thermal fraction in the lmc of 0.85 .
this value is probably an underestimate due to the fact that , while @xmath99 is a good approximation for frequencies less than or equal to a few ghz and @xmath100 , the actual free - free spectral index is a weak function of both frequency and electron temperature @xcite .
moreover , hii regions span a wide range of t@xmath49 values .
if we compare our result with similar attempts found in the literature , we see that , for instance , @xcite tried to separate the thermal and non thermal component in the radio emission of the lmc , using h@xmath2 and fir data , at a resolution of 51@xmath17 . at 4.75 ghz , they obtained a thermal fraction of 0.59 .
however their result applies to both the diffuse medium and discrete hii regions , while we only consider hii regions .
+ in order to take the synchrotron contamination into account , we multiply the 4.75-ghz radio map by the derived thermal fraction equal to 0.85 .
+ results of the correlations for each phase and regime of the gas are presented in figure [ fig_allcases ] and in table [ table_val ] . from equation [ eq_gen ]
we obtain the emissivities associated with each phase of the gas for each wavelength @xmath101 in the 3.6 @xmath0 - 160@xmath0 range , and with these values we build the corresponding spectral energy distributions ( seds ) . the seds are modeled with _ dustem _ @xcite .
the _ dustem _
code is an updated version of the @xcite model .
the code allows to constrain mass abundances ( @xmath102 , @xmath103 , and @xmath104 ) , relative to hydrogen , of the three dust populations .
the intensity of a nir continuum , with a temperature of 1000 k , is also a free parameter of the model .
indeed , the inclusion of this component is required , as discussed in several studies ( e.g. * ? ? ?
* ) , and allows us to reproduce observational data at 3.6 and 4.5 @xmath0 .
two different radiation fields ( rfs ) input models are used in this work .
the strength of the field ( @xmath105 ) is a free parameter in both cases .
however , for all 3 cases ( diffuse medium , typical and very bright hii regions ) , we impose , for dust emission associated with the atomic and molecular phases , as well as for dust emission associated with the diffuse ionized gas in case 1 , the same interstellar rf hardness as for the solar neighborhood @xcite . in the following
we will refer to this rf as _ mathis _
we note that , for dust in the atomic and molecular gas phases , we assume that the dominant component of the rf is always the one related to the interstellar solar neighborhood .
however for cases 2 and 3 , dust in the ionized phase is associated with typical and bright hii regions , and therefore it is essentially heated by local hot stars . in these regions , the solar neighborhood
interstellar rf is not a good approximation .
following @xcite , the youngest stellar objects are less than 10 myr . using the galev code , which is an implementation of evolutionary synthesis models
, we have generated an uv / visible spectrum of young stellar clusters of 4 myr . in the following
, we will refer to this rf as the _
4-myr cluster _ rf .
the summary of the rfs used for each case and phase is described in table [ table_cases ] , and the rf spectral shapes are shown in figure [ fig_isrf ] . assuming a single rf component along the los is a simplistic hypothesis .
therefore we tested our results using instead a combination of several rf intensities and hardnesses along the los .
for case 1 , we apply the `` local - sed combination '' model proposed by @xcite for normal star forming galaxies .
this model assumes a power - law distribution of dust mass subject to a given heating intensity @xmath106 : @xmath107 where @xmath108 controls the relative contribution of the various strengths of the field to the seds . @xcite
considered rf strengths and @xmath108 values in the range @xmath109 and @xmath110 , respectively .
a value of @xmath2 close to 1 indicates an active star - forming region , whereas @xmath2 around 2.5 corresponds to the diffuse medium .
this parameter is linked to the i@xmath111/i@xmath112 ratio , and therefore gives an indirect measure of dust temperature . + here
is how we proceed in our analysis .
we first compute emission spectra using _ dustem _ , for different @xmath113 values , and the _ mathis _ rf ( noted @xmath114 ) .
then we sum up these emission spectra over the same @xmath105 range as proposed by @xcite , according to : @xmath115 for the atomic and molecular phases in cases 2 and 3 , we check the impact of adopting a combination of the _ mathis _ rf and the simulated _ 4-myr cluster
_ galev rf .
equation [ eq_isrf ] becomes : @xmath116 the coefficient @xmath117 allows to normalize the emission spectra computed with each rf hardness .
indeed , the same @xmath113 will not give the same absolute emission level depending on the rf used . to determine @xmath117 , we adjust the emission spectra using the two rfs ( _ mathis _ rf and _ 4-myr cluster _ rf ) and look for the coefficient allowing the same brightness at the maximum emission .
this coefficient @xmath117 is re - derived for each @xmath113 . for the ionized phase in cases 2 and 3
, we apply equation [ eq_isrf ] , but replacing the _ mathis _ rf by the _
4-myr cluster _ rf .
the _ dustem _ code does not allow to perform a model fit using a combination of rfs .
therefore , we have to use a different method to analyze the effect of a composite rf .
following @xcite , in order to derive mass abundances for each dust grain species , we pre - compute the ir brightnesses associated with each gas phase , using the galactic abundances obtained by @xcite , for different @xmath108 values .
we then compare the observed brightnesses between 8 and 160 @xmath0 with the predicted values , and infer the dust mass abundances minimizing the @xmath118 for each value of @xmath2 .
we select the solution with the lowest @xmath118 .
the additional nir continuum is not taken into account to compute the ir brightnesses , since we do not know the origin of this component , nor how to treat it in the case of a combination of rfs .
for this reason , we do not consider ir data below 8 @xmath0 .
the summary of the rf combinations used for each case and phase is given in table [ table_cases_sum ] .
the dust emission spectra are shown in figure [ fig_allcases ] .
figure [ fig_model ] illustrates the sed modeling obtained using _ dustem _ , assuming a single radiation field component . in the following
we analyze the results for the diffuse ionized gas , typical and bright hii regions .
we also discuss their consistency using the hypothesis of a rfs combination .
the best - fit parameters for each phase of the gas ( molecular , atomic and ionized ) , and each regime of the ionized gas , as well as for different rfs assumptions , are provided in tables [ table_cases ] and [ table_cases_sum ] .
figures [ fig_allcases ] ( panel a ) and [ fig_model ] indicate that , when compared to the atomic and molecular phases , dust emission associated with the ionized gas phase is a relevant contribution to the total emission . however , we emphasize that the absolute amplitude of the spectrum for the ionized phase is proportional to n@xmath49 . as a consequence ,
if n@xmath49 has been underestimated / overestimated , the absolute level of the spectrum has also been underestimated / overestimated , although the spectral shape will remain unchanged . here , as discussed in section 3.1 , we take n@xmath49 equal to 0.055 @xmath86 , which is the average observed value in our galaxy , as reported by @xcite .
+ in this case , the adopted rf hardness is the same for all gas phases , as discussed in section [ sec_mod_single ] .
therefore , it is possible to perform a direct comparison among the temperature estimates resulting from the best - fit rf intensities .
we remind the reader that , if we assume a greybody dust emission model with spectral index @xmath119 , then the rf intensity and the equilibrium temperature of the bgs are related through : @xmath120 where x@xmath121 is the rf scaling factor for the _ mathis _ rf ( x@xmath122 ) , and the dust temperature in the solar neighborhood @xmath123 is taken equal to 17.5 k @xcite , with @xmath124 .
we find dust temperatures equal to 18.7 k , 16.1 k and 23.9 k for the atomic , molecular and ionized phases , respectively .
these results confirm that dust associated with the ionized phase is warmer than dust in the atomic and molecular gas phases , although our derived temperature for the diffuse h@xmath67 is slightly lower than the value obtained by @xcite , of 26.6 @xmath83 0.1 k. as for dust emissivity for the ionized phase , we obtain an estimate of 2.3@xmath125 @xmath32/h at 160 @xmath0 . for comparison ,
@xcite find , at high latitude , an emissivity of 1.1@xmath126 @xmath32/h at 250 @xmath0 , which corresponds to an emissivity of 2.7@xmath126 @xmath32/h at 160 @xmath0 assuming a spectral index equal to 2 , while @xcite , for the galactic plane , report a value of 35.1 mjy / sr for 10@xmath127 h/@xmath32 at 140 @xmath0 , corresponding to 8.7 @xmath126 @xmath32/h at 160 @xmath0 .
therefore , even if in terms of temperature , dust in the ionized phase of the lmc is roughly comparable to its galactic plane counterpart , it appears to be 40 times less emissive in the lmc with respect to the galactic plane , and 10 times less emissive with respect to the high galactic latitudes
. the low value in the lmc can not be only due to the low metallicity of the lmc ( z @xmath12 0.3 - 0.5 z@xmath128 * ? ? ? * ) with respect to our galaxy , as the discrepancy in emissivity is significantly larger than the difference in metallicity .
for the galactic plane , at least part of the difference could be induced by confusion along the los . however
, we think that different dust properties in the ionized gas phase of the lmc compared to our galaxy ( high latitudes and galactic plane ) are likely at the origin of the observed discrepancy . + from table [ table_cases ] , the total amount of dust in the ionized phase , similar to the dust - over - gas ratio ( d / g ) , appears to be 3.6 and 6.4 times lower than in the atomic and molecular phases . in particular , the pah relative abundance with respect to bgs reveals a decrement by a factor 2.9 and 6.6 when compared to the atomic and molecular phases , respectively .
the high value of pah relative abundance in the molecular phase is connected to the low x@xmath129 which , in turn , is a consequence of the shielding effect from the rf induced by the higher density characterizing this phase of the gas . for the vsg relative abundance in the ionized phase
, we do not find the same depletion behaviour as for pahs .
the vsg relative abundance is in fact roughly the same in the three phases .
we interpret this effect as vsgs being resiliant to the rf .
remarkably , the spectrum of dust associated with the ionized phase highlights more emission in the nir domain , by a factor of 7.5 and 18.5 , compared to the molecular and atomic phases .
the origin of this effect is not clear .
@xcite argue that the nir continuum intensity is quite linear with the strength of the aromatic features in emission .
they suggest that both this continuum and the aromatic features originate from similar carriers .
@xcite , using spitzer data for the galactic diffuse interstellar medium , require the presence of a continuum to explain the observations , but they conclude that the continuum carriers might not be the pahs .
what we observe in the ionized phase of the lmc is a decrease of the pah relative abundance and an increase of the nir continuum , whereas the low value of the nir continuum intensity in the molecular phase is associated with a high value of pah relative abundance .
therefore our results favour the @xcite s interpretation . in this case
, the modeling seems to reveal ( figure [ fig_allcases ] , panel b ) that emission associated with the ionized gas phase is not only a major contributor to the total emission , but is even dominant with respect to the other components , i.e. atomic and molecular . nevertheless , as in the previous section , we need to remind the reader that this result quite strongly depends on the electron density . at the same time
, we do not think that the adopted value can be off by a large amount , given that it has been derived from the hii regions in the hk86 catalog .
in fact , we should be wrong by a factor 5 - 6 in order for the ionized gas spectrum to match the atomic and molecular spectra at 160 @xmath0 . for the ionized phase associated with typical hii regions , we adopt - as discussed in section [ sec_mod_single ] - a bluer rf input model .
indeed , the _ mathis _
rf is not representative in this case , as the ionized gas is mostly heated by radiation from nearby stars . since we assume different rfs for different gas phases , we can not directly compare the temperature inferred from the model .
however , figure [ fig_model ] evidences a shift of the bg emission peak towards shorter wavelengths in the ionized gas phase compared to the other phases .
this result would suggest an increase of dust temperature when going from the atomic or molecular phase to the ionized one .
the hardness of the rf modeled with galev is proportional to @xmath130 .
we have checked our results against different slopes of the input rf , for example @xmath131 and @xmath132 ( see figure [ fig_isrf ] ) , and verified that the changes in the outputs are negligible .
as for case 1 , the ionized phase appears to be characterized by less pahs relative to bgs . in particular , the pah relative abundance decreases by a factor 4.5 to 10 , compared to the relative abundance in the atomic and molecular phases .
this is likely a consequence of a higher rf intensity with respect to the other phases .
furthermore , in a similar fashion to what we find for the diffuse ionized medium , the molecular phase presents an enhancement of the pah relative abundance . for typical hii regions ,
the co data trace the photo - dissociation region ( pdr ) surrounding the ionized gas , where the local high density contributes to shield the pahs from the stellar rf .
interestingly , the atomic phase also shows a decrease in the pah relative abundance compared to the molecular phase , even if the effect is less important than for the ionized phase .
the atomic phase is subject to an attenuated stellar rf , as well as to moderate radiation from the ism .
the combined effect of both rfs might cause the partial pah destruction , or a decrease of the excitation state of the pahs . in this latter scenario
, the pahs would be less emissive .
overall , the results obtained for typical hii regions confirm those derived for the diffuse h@xmath67 .
the vsg relative abundance is qualitatively the same in the three phases . as for case 1 , we notice that the nir continuum intensity is increased in the ionized phase with respect to the atomic phase , in this case by a factor 90 , whereas the pah relative abundance decreases .
this result corroborates the hypothesis that the nir continuum is not correlated with the pah abundance . in case 3
, we are looking at dust emission in the vicinity of conspicuous massive star associations .
as in the previous case , dust emission in the ionized phase appears to dominate the total emission ( see figure [ fig_allcases ] , panels c and d ) , although with the caveat discussed in section 4.1.1 and 4.1.2 .
results derived using h@xmath2 and radio data are in good agreement .
the lower amplitude of the spectrum associated with the ionized phase when in equation [ eq_gen ] we make use of the parkes rather than the shassa data is likely partly a consequence of dust absorption in the shassa map , as already mentioned in section [ sec_halpha_tracers ] . from table
[ table_cases ] , we notice that the rf intensity and the dust relative abundances are in good agreement for both the ionized gas tracers .
the results show , for the ionized gas phase in particular , a low pah relative abundance comparable to case 2 , and an increase of the vsg relative abundance by a factor 2.2 - 2.5 ( @xmath133 in case 2 and @xmath134 in cases 3 ) . remarkably , these conclusions do not depend on the specific spectral shape of the rf modeled with galev .
the molecular phase appears again to be characterized by a high pah relative abundance .
we note that the pah destruction in the ionized phase becomes more pronounced going from case 1 to case 3 , probably due to the increase of the rf intensity from the diffuse medium to bright hii regions .
this is consistent with previous findings reported in the literature .
for instance , using isocam data , @xcite found a lack of aromatic features carriers in the core of the lmc hii region n4 , and suggested that the high and hard rf might be responsible for the pah destruction .
likewise , @xcite , identified pah destruction at the edge of the hii region m17 .
@xcite explained the decrease of the 12 @xmath0 emission in the california nebula by grain destruction in regions with high uv radiation .
they estimated that around 80@xmath69 of the 12 @xmath0 emitters could be destroyed whenever the energy density is larger than 50 times the energy density in the solar neighborhood . despite this theoretical and observational evidence
, we can not state that the rf is systematically responsible for the depletion of pahs in hii regions .
indeed , several other processes could be at work , like coulomb destruction of pah dications or dehydrogenation by direct photodissociation , as suggested by @xcite based on the analysis of the 3.3 @xmath0 emission and nearby continuum in the planetary nebula bd+30 3639 .
moreover , @xcite studied the same dust feature in the orion bar and m17 , and propose that h@xmath67 chemical sputtering could cause pah destruction . + in terms of temperature , as in case 2 , we can deduce from figure [ fig_model ] the presence of warmer dust in the ionized gas as compared to the other phases , from the fact that the peak of emission for the bgs is located at shorter wavelengths .
finally , as for case 1 and 2 , the nir continuum is higher in the ionized phase than in the other phases . in this section ,
we discuss the use of a combination of rfs , instead of a single rf , for fitting the seds corresponding to each phase of the gas ( see table [ table_cases_sum ] ) . for case 1 , the values given in table [ table_cases_sum ] confirm the decrease of the pah relative abundance in the ionized phase , as evidenced in section [ subsection_case1 ] . on the contrary
, the enhancement of the pahs in the molecular phase might suggest that these environmental conditions are indeed particularly favourable for their survival .
for the vsg relative abundance in the ionized phase , we find a significantly lower value with respect to what we obtain by adopting a single rf component ( see tables [ table_cases ] and [ table_cases_sum ] ) .
however , the vsg relative abundance is highly dependent on the quality of the fit at fir wavelengths and , in the ionized phase , the model exhibits an important lack of emission at 70 @xmath0 , as well as a higher @xmath118 , as compared to a single rf modeling .
this might indicate that the low vsg relative abundance has , for a composite rf , been underestimated . for case 2 , the combination of the _ mathis _ with the _ 4-myr cluster _ rfs , both for the atomic and molecular phases , generates a bluer rf than a single _ mathis _ rf .
therefore , as expected , we find a decrease in the pah and vsg relative abundances .
indeed , pahs and vsgs absorb most of their energy at short wavelengths , while bgs absorb energy over the whole rf spectrum . as a consequence , pahs and vsgs
absorb and re - emit more energy with a bluer rf , and to produce the same ir flux with a bluer rf than the _ mathis _
rf , the model requires less pahs and vsgs . for the ionized phase , the combination of _
4-myr cluster _ rfs gives the same pah relative abundance of a single _ 4-myr cluster _ rf , whereas a significant decrease of the vsg relative abundance is observed .
however , the general trend is similar to what found for a single rf : as in case 1 , the highest pah relative abundance is in the molecular phase , which also exhibits more vsgs with respect to bgs , while the atomic phase shows intermediate values of pah and vsg relative abundances , compared to the other gas phases . in case 3 , as for case 2 , the pah relative abundance in the atomic and molecular phases decrease by a factor @xmath3 2 compared to a single rf . in the ionized phase , we still have less pahs than in the atomic phase , by a factor 2.2 and 2.7 , for case 3a and 3b , respectively , and by a factor between 4.4 and 5.1 compared to the molecular phase . results
are therefore in agreement with the modeling with a single rf . for the vsg relative abundance , we obtain , for the atomic phase , different and contradictory results depending on the assumption on the rf , therefore preventing a clear interpretation of the results . in the ionized phase ,
the modeling confirms the enhancement of the relative vsg abundance already highlighted by a single rf .
overall , the results obtained assuming a composite rf substantially corroborate those derived with a single rf . in panels
e , f and g of figure [ fig_allcases ] , we present the results of the analysis in a slightly different fashion , i.e. by combining in individual plots , each corresponding to a given gas phase , the seds obtained for all cases ( 1 , 2 and 3 ) .
we can see that no matter which gas component we consider , the amplitude of the spectrum always increases going from the diffuse medium to bright hii regions .
however , for the atomic phase , the difference in amplitude between the different cases is less important compared to the other gas components , and a rather remarkable agreement among cases is found at 8 @xmath0 . as previously noted , the absolute level of dust emission in the ionized gas is directly proportional to @xmath46 , for which the adopted value changes according to the specific environment .
therefore , the discrepancy in absolute amplitude of the spectra for that phase is not well constrained . on the contrary , for the atomic and molecular phases ,
we are confident that the recovered sed amplitudes are reliable , as these depend on the hi and co conversion factor ( equation 1 and 2 ) which , respectively , is known to be constant for a given environment ( hi ) , and has been accurately determined for the majority of the clouds ( co ) .
we observe prominent 8 @xmath0 emission in the atomic phase , probably associated with the 7.7 @xmath0 pah feature .
this emission feature does not show in the ionized phase for the case of hii regions , likely due to pah depletion . in comparison ,
the molecular phase is clearly characterized by strong 8 @xmath0 emission , as well as by significant emission at 5.8 and 12 @xmath0 . in the fir domain
, the atomic phase presents an excess of emission at 70 @xmath0 , as does the ionized phase in the diffuse medium .
this finding is in agreement with @xcite .
combining ir data with tracers of hi , h@xmath31 and h@xmath67 , we have evidenced , for the first time , dust emission associated with the ionized gas in the lmc .
we have performed our analysis for different regimes of the ionized gas : diffuse gas ( case 1 ) , typical hii regions ( case 2 ) and compact / bright hii regions ( case 3 ) .
we have modeled the spectra associated with the atomic , molecular and ionized phases of the gas , using different assumptions for the radiation field , to test the robustness of our results .
the results obtained by a combination of radiations fields along the line of sight confirm the results derived by assuming a single radiation field .
we report a systematic warmer dust temperature in the ionized phase compared to the atomic and molecular phases .
the inferred emissivity in the diffuse ionized gas is 2.3@xmath125 @xmath32/h at 160 @xmath0 , which is lower than the galactic values by a factor higher than the metallicity ratio between these 2 galaxies .
this result might suggest different properties of dust in the ionized gas of the lmc , compared to our galaxy .
we also find , for all cases ( 1 , 2 , 3 ) , a significant decrease of the pah relative abundance in the ionized phase compared to the other phases , with a larger difference between the phases for cases 2 and 3 .
we interpret this result as due to pah destruction , probably caused by the increased radiation field in the ionized phase , although the origin of this phenomenon is still under investigation . at the same time
, the molecular phase appears to favor the survival of the pahs .
in addition , when one compares bright hii regions with typical hii regions , the ionized phase shows an enhancement of the vsg relative abundance by more than a factor 2 .
we also find an important increase of the nir continuum in the ionized phase with respect to the other gas phases , which does not seem to correlate directly with pah emission . by comparing results derived using the h@xmath2 shassa map and the parkes radio data , the estimated extinction in h@xmath2 for bright hii regions
is found equal to 0.51 mag ( or a(v)=0.63 mag ) .
the origin of the destruction of the pahs in the ionized phase ( especially , in hii regions ) is still not fully understood .
moreover , the depletion of this species is not a systematic effect , suggesting that changes in the pah properties could also occur from one hii region to another . a forthcoming publication will investigate pah depletion by means of the spectroscopic spitzer / irs data available in the context of the sage - spec legacy survey @xcite . this work will allow us to study variations of the pah emission at small angular scale within individual hii regions .
we thank the referee for her / his useful comments , which helped to improve the content of the paper .
we also thank m. filipovic and a. hughes for helpful discussions on the parkes radio data and on the derivation of the thermal fraction .
we acknowledge the use of the dustem package . the nanten project is based on a mutual agreement between nagoya university and the carnegie institution of washington ( ciw ) . this work is financially supported in part by a grand - in - aid for scientific research from jsps ( no .
22244014 and no .
30377931 ) .
bennett , c. l. , et al .
1992 , apj , 396 , l7 bennett , c. l. , et al .
2003 , apj , 583 , 1 benot , a. , et al .
2002 , astropart . phys . , 17 , 101 bernard , j .- p . , giard , m. , normand , p. , @xmath135 tiphne , d. 1994 , a@xmath135a , 289 , 524 bernard , j .-
, et al . 2008 , apj , 136 , 919 boulanger , f. , et al . 1988 , apj , 332 , 328 boulanger , f. , abergel , a. , @xmath135 bernard j .-
1996 , a@xmath135a , 312 , 256 compigne , m. , et al .
2008 , a@xmath135a , 491 , 797 compigne , m. , et al .
2011 , a@xmath135a , 525 , 103 contursi , a. et al .
2007 , a@xmath135a , 469 , 539 dale , d. a. , helou , g. , contursi , a. , et al .
2001 , apj , 549 , 215 davies , r. d. , elliott , k. h. , @xmath135 meaburn , j. 1976 , mnras , 81 , 89 dsert , f .- x . , boulanger , f. , @xmath135 puget , j .- l .
1990 , a@xmath135a , 237 , 215 dickinson , c. , davies , r. d. , and davis , r. j. 2003 , mnras , 341 , 369 dobashi , k. , et al .
2008 , a@xmath135a , 484 , 205 everett , j. e. , @xmath135 churchwell , e. 2010 , apj , 713 , 592 feast , m. 1999 , in iau symp .
190 , new view of the magellanic clouds , ed .
h . chu et al .
( san francisco : asp ) , 542 filipovic , m. d. et al . 1995 ,
a@xmath135as , 111 , 311 flagey , n. et al .
2006 , a@xmath135as , 453,969 fukui , y. , et al .
2008 , apjs , 178 , 56 gaustad , j. e. , mccullough , p. r. , rosing , w. , @xmath135 van buren , d. 2001 , pasp , 113 , 1326 giard , m. et al .
1994 , a@xmath135a , 291 , 239 giardino , g. , et al .
2002 , a@xmath135a , 387 , 82 gordon , k. d. , et al .
2003 , apj , 594 , 279 groves , b. , nefs , b. , @xmath135 brandl , b. 2008 , mnras , 391 , 113 haffner , l. m. , et al .
2009 , rvmp , 81,969 hauser , m. g. 1993 , back to the galaxy , ed . s. s. holt , @xmath135 f. verter ( new - york : aip ) , aip procc . , 278 , 201 hoare , m. g. , roche , p. f. , @xmath135 glencross , w. m. 1991 , mnras , 251 , 584 hoernes , p. , berkhuijsen , e. m. , @xmath135 xu , c. 1998 , a@xmath135a , 334 , 57 hughes , a. , et al .
2010 , mnras , submitted hughes , a. , et al .
2006 , mnras , 370 , 363 imara , n. , @xmath135 blitz , l. , 2007 , apj , 662 , 969 kawamura , a. , et al .
2009 , apjs , 184 , 1 kemper , f. , et al . , 2010 , pasp , 122 , 683 kennicutt , r. c. , @xmath135 hodge , p. w. 1986 , apj , 306 , 130 kim , s. , et al .
2003 , apjs , 148 , 473 kotulla , r. , fritze , u. , weilbacher , p. , @xmath135 anders , p. 2009
, mnras , 396 , 462 lagache , g. , et al . 1999 ,
a@xmath135a , 344 , 322 lagache , g. , haffner , l. m. , reynolds , r. j. , @xmath135 tufte , s. l. 2000 , a@xmath135a , 354 , 247 van leeuwen , f. et al .
2007 , mnras , 379 , 723 lu , n. , et al .
2003 , apj , 588 , 199 mathis , j. s. , mezger , p. g. , @xmath135 panagia , n. 1983 , a@xmath135a , 128,212 meinert , d. , et al .
1993 , lnp , 416 , 130 meixner , m. , et al .
2006 , apj , 132 , 2268 miville - deschnes , m. a. , @xmath135 lagache , g. 2005 , apjs , 157 , 302 neugebauer , g. , et al .
1984 , apj , 278 , l1 niklas , s. , klein , u. , @xmath135 wielebinski , r. 1997 , a@xmath135a , 322 , 19 odonnell , j. e. 1994 , apj , 422 , 158 odegard , n. , et al .
2007 , apj , 667 , 11 padoan , p. , kim , s. , @xmath135 goodman , a. 2001 , apj , 555 , l33 paladini , r. , de zotti , g. , davies , r. , @xmath135 giard , m. 2005 , mnras 360 , 1545 paladini , r. , et al .
2007 , a@xmath136 465 , 839 paradis , d. et al . 2009a , aj , 138 , 196 paradis , d. et al .
2009b , a@xmath135a , 506 , 745 paradis , d. et al .
2011 , aj , 141 , 43 peeters , e. et al .
2005 , apj , 620 , 774 planck collaboration 2011e , submitted to a@xmath135a , arxiv1101.2029 povich , m. s. , et al .
2007 , apj , 660 , 346 reynolds , r. j. , tufte , s. l. , haffner , l. m. , et al . 1998 ,
pasa , 15 , 14 rieke , g. h. , et al .
2004 , apjs , 154 , 25 spitzer , l. 1978 , physical processes in the interstellar medium , wiley - interscience , new - york staveley - smith , l. , et al .
2003 , mnras , 339 , 87 stepnik , b. , et al .
2003 , a@xmath135a , 398 , 551 watson , c. , et al .
2009 , apj , 694 , 546 westerlund , b. e. 1997 , _ the magellanic clouds _ , new york , cambridge univ .
press wilson , t. l. , mezger , p. g. , gardner , f. f. , @xmath135 milne , d. k. 1970 , a@xmath135a , 6 , 364 | a systematic investigation of dust emission associated with the ionized gas has so far been performed only in our galaxy and for wavelengths longer than 60 @xmath0 .
newly available spitzer data now offer the opportunity to carry out a similar analysis in the large magellanic cloud ( lmc ) . by cross - correlating spitzer sage ( surveying the agents of a galaxy s evolution ) data with the atca /
parkes hi 21-cm data , the nanten @xmath1 ( j=1 - 0 ) data , and both the shassa h@xmath2 and the parkes 6-cm data , we investigate the physical properties of dust associated with the different phases of the gas ( atomic , molecular and ionized ) .
in particular , we study the presence and nature of dust from 3.6 to 160 @xmath0 and for various regimes of the ionized gas , spanning emission measures ( em ) from @xmath3 1 pc @xmath4 ( diffuse component ) to @xmath3 10@xmath5 pc @xmath4 ( hii regions ) . using a dust emission model , and testing our results with several radiation field spectra ,
we show that dust in the ionized gas is warmer than dust associated with other phases ( atomic and molecular ) .
we also find a decrease of the polycyclic aromatic hydrocarbons ( pah ) relative abundance with respect to big grains ( bgs ) , as well as an increase of the near infrared ( nir ) continuum .
these three results ( e.g. warmer temperature , decrease of pah abundance and increase of the nir continuum ) are found consistently for all regimes of the ionized gas . on the contrary
, the molecular phase appears to provide favorable conditions for the survival of pahs .
furthermore , the very small grain ( vsg ) relative abundance tends to increase in the ionized phase , especially in bright hii regions .
last but not least , our analysis shows that the emissivity of dust associated with the ionized gas is lower in the lmc than in our galaxy , and that this difference is not accounted for by the lower metallicity of the lmc . |
the unpredictability of turbulence makes a deterministic analysis of the instantaneous velocity field not only impractical , but very nearly impossible .
researchers have instead studied the statistical properties of turbulence , which necessarily involves the probabilities of velocities and their fluctuations @xcite .
although information theory is the natural language for treating these probability distributions @xcite , there have been few studies that make use of it . instead
, the focus has often been on the moments @xcite . in wall - bounded flows , for example
, considerable effort has been directed towards determining the mean velocity profile as a universal function of distance from the wall @xcite . in other situations ,
the fluctuations are of primary interest and the focus has been on the moments of velocity differences @xcite : @xmath0 ^n \rangle_x.\ ] ] these velocity differences are thought to represent the characteristic velocity of a turbulent eddy " of size @xmath1 , a concept that goes back to l.f .
richardson and his contemporaries @xcite .
the importance of @xmath2 in turbulence is apparent from its appearance in kolmogorov s @xmath3th law for @xmath4 @xcite .
this law is derived using several significant assumptions and the navier - stoke s equations @xcite and remains one of the only exact solutions in turbulence .
it is the starting point for the entire scaling phenomenology of turbulence . here
we propose a different approach . instead of beginning with the moments @xmath5 and kolmogorov - type assumptions , we will focus on the probability distributions used to calculate the @xmath5 , @xmath6 @xcite : @xmath7 with these probability distributions
, information theory can be used to make quantitative statements about , @xmath8 , the unpredictability of @xmath2 .
of course , all the information about the moments is contained in @xmath6 , so our analysis can not be completely unrelated to the traditional theory . in order to gauge the usefulness of this approach ,
we apply it to a key turbulence concept .
big whirls have little whirls + that feed on their velocity , + and little whirls have lesser whirls + and so on to viscosity .
+ the cascade concept captured by this rhyme can be formulated with information theory , because the @xmath9 of the cascade leaves its mark on the probability distributions of the eddies ( whirls ) . in short , the uncertainty in observing a big eddy become a small eddy should be less than the reverse . using experimental observations of ( quasi-)2d turbulent flow ,
we confirm the eddy hypothesis without the navier - stoke s equations , scaling arguments or kolmogorov s assumptions , although some of the features of that kind of analysis ( self - similarity @xcite ) reappear naturally .
a completely new result is the existence and direction of information transfer . before we can come to these conclusions , however
, we must review the salient features of turbulence and information theory .
two - dimensional ( 2d ) turbulence occurs approximately in nearly all large - scale atmospheric flows due partly to the fact that the thickness of the earth s atmosphere is very small compared to its breadth @xcite .
this , along with stratification @xcite , result in vast regions of the atmosphere where the vertical velocity is negligible compared with the horizontal .
for the same reason large scale oceanic flows are also considered two - dimensional .
our measurements are made using a soap film , which has an even smaller aspect ratio .
the physics of soap films and their usefulness in studying 2d flows and 2d turbulence has already been well documented @xcite , but we present some of the essential experimental details below . although we utilize 2d turbulence , our results should extend to 3d turbulence without any significant alteration .
the soap solution is a mixture of dawn ( 2@xmath10 ) detergent soap and water with 10 @xmath11 m hollow glass spheres added for the velocity measurements .
figure [ setup ] is a diagram of the experimental setup .
the soap film is suspended between two vertical blades connected to a nozzle above and a weight below by nylon fishing wire .
the nozzle is connected by tubes to a valve and a reservoir which is constantly replenished by a pump that brings the spent soap solution back up to the reservoir .
there is an overflow that leads back to the bottom reservoir so that the height of the top reservoir , and thus the pressure head , is constant
. sometimes the pump feeds directly into nozzle , giving results indistinguishable from the reservoir setup .
the flow is always gravity - driven .
typical centerline speeds @xmath12 are several hundred cm / s with rms fluctuations @xmath13 generally on the order of 10 cm / s .
the channel width @xmath14 is usually several cm .
the flow velocity is measured using particle image velocimetry ( piv ) @xcite .
a bright white light source is placed behind the soap film and shines on the soap film at an angle ( so that the light does not directly enter the camera , which is perpendicular ) .
the particles scatter more light than the surrounding soap water and are easily distinguished .
a fast camera tracks their movement and standard piv techniques are used to calculate the velocity field . because the full velocity field at different times is needed for our analysis , we use piv and can not use , @xmath8 , laser doppler velocimetry .
we focus on the wall - normal or horizontal velocity @xmath15 instead of the vertical velocity @xmath16 because it does not suffer from aliasing effects @xcite and because @xmath12 changes slightly with @xmath17 ( due to gravitational acceleration ) .
we take velocities very near the center of the channel to avoid wall effects and variations in the energy and enstrophy ( norm of vorticity ) injection rates @xcite .
turbulence in the soap film is generated by inserting a row of rods ( comb ) perpendicular to the film .
when this protocol is used we almost always observe the direct enstrophy cascade @xcite . in the traditional phenomenology
, enstrophy is transferred from some large injection scale @xmath18 @xmath19 to a viscous dissipative scale @xmath20 @xcite .
the rate of enstrophy transfer is @xmath21 , which should be constant over the inertial range of scales in between @xmath18 and @xmath20 .
we introduce this traditional framework in order to contrast it with our own approach , as well as provide evidence to the portential skeptic that we are indeed working with a direct cascade .
as is customary we assume that the energy spectrum @xmath22 of velocity fluctuations scales with @xmath21 and the wavenumber @xmath23 , an inverse length , for a certain range of @xmath23 ( inertial range ) . for the direct enstrophy cascade , @xmath24 . in fig .
[ spectra ] we plot all of the @xmath22 calculated from @xmath15 .
the curves have been normalized using the large length scale @xmath25 and the rms velocity @xmath26 .
all of the curves collapse at low and intermediate @xmath23 , a signature of the cascade s self - similarity @xcite .
the physical process that created this curve is similar in all cases .
we now outline our test .
consider the diagram of a turbulent cascade shown in fig .
[ cascade ] , which is typical of that shown in many turbulence textbooks @xcite .
the richardson picture involves eddies evolving in both space and time .
a large eddy at an early time ( @xmath30 ) will become a small eddy at a later time ( @xmath31 ) .
( we also consider the reverse process . )
we should be more certain about @xmath31 given @xmath30 than @xmath32 given @xmath33 .
let us now move on to the precise description .
the shannon entropy is central to information theory @xcite .
we could simply look at the raw probability distributions themselves , but the entropy gives a single number that characterizes how random the distribution is and provides an interpretation in terms of information .
for the eddy @xmath34 , the shannon entropy is @xmath35 where the sum is over all possible values of @xmath33 .
this is the amount of information we gain from or uncertainty we had prior to measuring @xmath33 @xcite .
( uncertainty and information are the same in this framework . ) to test a relationship between two eddies , we use a modified form of the entropy : the conditional entropy @xcite .
this gives us the uncertainty of one eddy given that the other eddy occurred . for the uncertainty of
@xmath31 given @xmath30 we write @xmath36 where @xmath37 and @xmath38 are the joint and conditional probabilities respectively .
if @xmath30 and @xmath31 are independent , @xmath39 . knowing @xmath30 does nt help us reduce our uncertainty about @xmath31 . if @xmath31 is determined by @xmath30 with absolute certainty , then @xmath40 .
so the stronger the relationship , the smaller this quantity is .
the @xmath41 take on a continuous range of values , but to calculate probabilities and then estimate entropies , we must bin ( discretize ) the measured data .
this is , in fact , unavoidable due to the finite resolution of all measurement apparatuses .
we systematically varied the bin size , but as in ref .
@xcite , we found our results to be extremely robust .
none of our results will change except for a vertical shift in all the conditional entropies by the same factor .
we use a bin size of @xmath42 , which of course changes size with @xmath1 , but ensures that the number of bins ( and thus the maximum possible value of the conditional entropy ) remains roughly the same .
we assert that @xmath43 signifies a cascade from large to small scales , and we use @xmath44 to test this .
@xmath45 in principle also depends on @xmath46 , @xmath47 and @xmath48 as indicated in fig .
[ cascade ] , but we maximize with respect to @xmath47 and @xmath48 ( experimental parameters ) and take many realizations at different @xmath46 , which is simply a reference marker , so that in the end @xmath45 only depends on @xmath1 .
this means that we are formulating richardson s hypothesis @xmath49 at a scale @xmath1 .
we note that our approach is similar to work on information transport in spatiotemporal systems @xcite . a study that more closely anticipates our own is an information transfer treatment of the goy model @xcite .
the result of the calculation of @xmath45 is shown in fig .
[ cond_entropy_a ] for the same @xmath27 and @xmath22 as in fig .
[ spectra ] . clearly @xmath50 is greater than zero in all cases as expected for a direct cascade . in the traditional framework ,
a cascade with constant transfer rate is first assumed and then the scalings are derived with the extra assumptions ( universality , etc . ) already mentioned @xcite .
then if the moments scale as predicted , the cascade is considered to be demonstrated . here
we bypass this sophisticated argumentation and show the cascade straightaway .
more can be gleaned from fig .
[ cond_entropy_a ] .
there appears to be a region of nearly constant value in many of the curves , which is reminiscent of the enstrophy ( energy ) flux which takes on a maximum and constant value in the inertial range equal to the injection rate @xmath21 ( @xmath51 ) @xcite .
this suggests a connection between max(@xmath52 ) and @xmath21 . estimating @xmath21 from fig .
[ spectra ] , we find a general increase of max(@xmath52 ) with @xmath21 .
moreover , all of the curves appear to have a similar shape .
this is reminiscent of the energy spectra above , and suggestive of an underlying self - similarity . indeed by choosing a single @xmath53 and normalizing @xmath1 by @xmath53 and @xmath45 by @xmath54 , we find reasonable collapse as seen in fig .
[ cond_entropy_b ] .
not only have we rediscovered self - similarity , but also a turbulent length scale .
@xmath53 is very close to @xmath18 , but their ratio decreases with @xmath27 as shown in the inset of fig .
[ cond_entropy_b ] .
now let us establish an interesting corollary .
the mutual information is the information shared between two variables . for the large eddies at earlier time ( @xmath30 ) and the small at later time ( @xmath31 ) , we write @xmath55 if we also define a mutual information for @xmath33 and @xmath32 , then we can rewrite @xmath52 as @xmath56 where we have used that @xmath57 , due to the small @xmath48 ( experimentally verified ) .
this means that @xmath58 in other words , the information shared between eddies going downscale is more than the reverse .
there is net information being transferred downscale , concurrently with the enstrophy .
this result should apply generally to both the 2d and 3d direct cascades and even the 2d inverse energy cascade .
it rests only on the validity of our information theory expression of richardson s eddy hypothesis , which we have here experimentally verified .
the implications of this result are quite powerful .
kolmogorov s small scale universality assumption can be expressed as the small scales forgetting " about the large ( forcing ) scales @xcite .
[ eq : mutinfo ] suggests that this can never be true .
so long as there is a cascade , there must be information transfer in the same direction which makes forgetting " impossible .
thus intermittency appears to be a necessary feature of all turbulent cascades .
we have shown that we can formulate and test richardson s idea of a cascade without using the navier - stoke s equation , scaling arguments or kolmogorov s assumptions .
and yet some of the old patterns , such as self - similarity , re - emerged .
this work presents an entirely new perspective on the statistics of turbulent velocities and suggests a more suitable framework for understanding intermittency .
our application of information theory to turbulence ought to serve as a guidepost for further work . | turbulence theory is usually concerned with the statistical moments of the velocity and its fluctuations .
one could also analyze the implicit probability distributions .
this is the purview of information theory .
here we use information theory , specifically the conditional entropy , to analyze ( quasi-)2d turbulence .
we recast richardson s eddy hypothesis " that large eddies break up into small eddies in time using the language of information theory .
in addition to confirming richardson s idea , we find that self - similarity and turbulent length scales reappear naturally .
not surprisingly , we also find that the direction of information transfer is the same as the direction of the cascade itself .
consequently , intermittency may be considered a necessary companion to all turbulent flows . |
this work was supported by the korea science and engineering foundation(kosef ) grant funded by the korea government(mest ) ( 2006 - 8 - 0083 ) .
999 c. f. von weizscker , z. phys .
* 96 * ( 1935 ) 431 .
e. ha and d. cha , j. korean phys
* 50 * ( 2007 ) 1172 .
d. kim , e. ha , and d. cha , nucl .
a799 * ( 2008 ) 46 . s. raman , c. w. nestor , jr . , and p. tikkanen , at .
data nucl .
data tables * 78 * ( 2001 ) 1 .
t. kibdi and r. h. spear , at .
data nucl .
data tables * 80 * ( 2002 ) 35 .
r. b. firestone , v. s. shirley , c. m. baglin , s. y. f. chu , and j. j. zipkin , _ table of isotopes _ , 8th edition ( wiley , new york , 1999 ) . b. sabbey , m. bender , g. f. bertsch , and p .- h .
heenen , phys . rev . *
c*75 ( 2007 ) 044305 .
p. d. cottle , phys .
* c42 * ( 1990 ) 1264 .
w. nazarewicz , p. olanders , i. ragnarsson , j. dudek , g. a. leander , p. mller , and e. ruchowska , nucl . phys .
* a429 * ( 1984 ) 269 . for a review of the octupole correlations in nuclei ,
see i. ahmad and p. a. butler , annu .
nucl . part .
* 43 * ( 1993 ) 71 .
ensdf database , offerred by the national nuclear data center at the brookhaven national laboratory , is available on the world wide web at http://www.nndc.bnl.gov / ensdf/. r. f. casten , phys .
* 54 * ( 1985 ) 1991 .
r. f. casten and n. v. zamfir , j. phys .
g : nucl . part .
* 22 * ( 1996 ) 1521 .
* figure 1*. the measured lowest excitation energies @xmath2 of the natural parity states in even - even nuclei @xcite .
the plotted points are connected by solid lines along the isotopic chains .
part ( a ) shows the lowest excitation energies @xmath2 of the even multipole states while part ( b ) shows those of odd multipole states except the dipole states . in part ( c ) , the lowest excitation energies @xmath2 of the dipole states are displayed on top of the same data shown in part ( b ) .
* figure 3*. the lowest excitation energies @xmath2 of the natural parity odd multipole states in even - even nuclei between @xmath59 and @xmath60 .
the filled triangles and circles denote the measured excitation energies while the empty ones show the excitation energies calculated by eq.([ge ] ) .
* figure 4*. the excitation energies @xmath2 of the lowest dipole states in even - even nuclei .
the upper part shows the measured excitation energies while the lower part shows those calculated by eq.([ge ] ) .
the measured excitation energies in the left upper panel are extracted from the table of isotopes , 8th edition by firestone _
et al_. @xcite , while the additional excitation energies , which are extracted from the ensdf database @xcite , are plotted in the right upper panel by open circles .
* figure 5*. the same as in fig.[fig-4 ] , but for the excitation energies @xmath2 of the lowest @xmath44 states in even - even nuclei .
the measured excitation energies in the left upper panel are quoted from the compilation in kib ' edi _ et al_. @xcite .
* figure 6*. the same as in fig.[fig-4 ] , but for the excitation energies @xmath2 of the lowest @xmath45 states in even - even nuclei .
the measured excitation energies in the left upper panel are extracted from the table of isotopes , 8th edition by firestone _
et al_. @xcite .
* figure 7*. the same as in fig.[fig-4 ] , but for the excitation energies @xmath2 of the lowest @xmath46 states in even - even nuclei .
the measured excitation energies in the left upper panel are extracted from the table of isotopes , 8th edition by firestone _
et al_. @xcite .
* figure 8*. the same as in fig.[fig-4 ] , but for the excitation energies @xmath2 of the lowest @xmath47 states in even - even nuclei .
the measured excitation energies in the left upper panel are extracted from the table of isotopes , 8th edition by firestone _
et al_. @xcite .
* figure 9*. the scatter plot of the calculated excitation energies @xmath48 as a function of the measured energies @xmath49 for the lowest excitation energies of the natural parity odd multipole ( left and middle panels ) as well as even multipole ( right panels ) states .
the scatter plots shown in the middle panels employ only the additional excitation energies extracted from the ensdf database @xcite .
the graphs for the even multipole states are quoted from ref.@xcite . * figure 10*. two of the scatter plots , particularly for the octupole states ( left panel ) among the left panels of fig.[fig-9 ] , and for the quadrupole states ( right panel ) among the right panels of fig.[fig-9 ] are redrawn .
the plotted points are expressed by the following different symbols : solid circles ( both of proton and neutron shells are closed ) ; solid squares ( only one of proton and neutron shells is closed ) ; open circles ( both of proton and neutron shells are not closed ) .
.the values adopted for the six parameters in eq.([ge ] ) for the lowest excitation energies @xmath2 of the natural parity including even multipole @xmath61 , @xmath62 , @xmath63 , @xmath64 , and @xmath65 states and also odd multipole @xmath43 , @xmath44 , @xmath45 , @xmath46 , and @xmath47 states .
the values for the even multipole states are quoted from ref.@xcite .
the last two columns are the @xmath26 value , defined by eq.([chi ] ) , which fits the parameter set and the total number @xmath28 of the data points , respectively , for the corresponding multipole state .
the numbers in the parentheses are obtained by including the additional excitation energies extracted from the ensdf database @xcite . [
cols="^,^,^,^,^,^,^,^,^ " , ] of the natural parity states in even - even nuclei @xcite .
the plotted points are connected by solid lines along the isotopic chains .
part ( a ) shows the lowest excitation energies @xmath2 of the even multipole states while part ( b ) shows those of odd multipole states except the dipole states . in part ( c ) , the lowest excitation energies @xmath2 of the dipole states
are displayed on top of the same data shown in part ( b).,width=491 ] of the natural parity odd multipole states in even - even nuclei between @xmath59 and @xmath60 .
the filled triangles and circles denote the measured excitation energies while the empty ones show the excitation energies calculated by eq.([ge]).,width=453 ] of the lowest dipole states in even - even nuclei .
the upper part shows the measured excitation energies while the lower part shows those calculated by eq.([ge ] ) .
the measured excitation energies in the left upper panel are extracted from the table of isotopes , 8th edition by firestone _
et al_. @xcite , while the additional excitation energies , which are extracted from the ensdf database @xcite , are plotted in the right upper panel by open circles.,width=529 ] , but for the excitation energies @xmath2 of the lowest @xmath44 states in even - even nuclei .
the measured excitation energies in the left upper panel are quoted from the compilation in kib ' edi _ et al_. @xcite.,width=529 ] , but for the excitation energies @xmath2 of the lowest @xmath45 states in even - even nuclei .
the measured excitation energies in the left upper panel are extracted from the table of isotopes , 8th edition by firestone _
et al_. @xcite.,width=529 ] , but for the excitation energies @xmath2 of the lowest @xmath46 states in even - even nuclei .
the measured excitation energies in the left upper panel are extracted from the table of isotopes , 8th edition by firestone _
et al_. @xcite.,width=529 ] , but for the excitation energies @xmath2 of the lowest @xmath47 states in even - even nuclei .
the measured excitation energies in the left upper panel are extracted from the table of isotopes , 8th edition by firestone _
et al_. @xcite.,width=529 ] as a function of the measured energies @xmath49 for the lowest excitation energies of the natural parity odd multipole ( left and middle panels ) as well as even multipole ( right panels ) states .
the scatter plots shown in the middle panels employ only the additional excitation energies extracted from the ensdf database @xcite .
the graphs for the even multipole states are quoted from ref.@xcite.,width=415 ] , and for the quadrupole states ( right panel ) among the right panels of fig.[fig-9 ] are redrawn .
the plotted points are expressed by the following different symbols : solid circles ( both of proton and neutron shells are closed ) ; solid squares ( only one of proton and neutron shells is closed ) ; open circles ( both of proton and neutron shells are not closed).,width=529 ] | we applied our recently proposed empirical formula , a formula quite successful in describing essential trends of the lowest excitation energies of the natural parity even multipole states , to the lowest excitation energies of the natural parity odd multipole states in even - even nuclei throughout the entire periodic table .
even though the systematic behavior of the lowest excitation energies of odd multipole states is quite different from those of even multipole states , we have shown that the same empirical formula also holds reasonably well for the odd multipole states with the exception of a few certain instances . empirical formula ;
lowest excitation energies of natural parity odd multipole states ; valence nucleon numbers + 21.10.re ; 23.20.lv nuclear physics can be explained from two standpoints .
one is the portrait description where the excited states of an individual nucleus are studied in depth .
the traditional nuclear structure models , which have contributed significantly to the development of nuclear physics , belong to this category .
the other is the landscape description where some particular nuclear properties are examined in terms of simple nuclear variables over the wide span of the chart of nuclides .
a well - known example of a landscape description is the weizscker s semi - empirical mass formula first proposed in the early 1930 s @xcite .
it can reproduce the binding energy of the ground state of nuclei very accurately throughout the entire chart of nuclides in terms of the mass number @xmath0 and the atomic number @xmath1 with just a few fitted parameters .
however , for a very long time , no expression of a similar sort has been available for the properties of excited states in nuclei until very recently when we proposed a simple empirical formula for the lowest excitation energies @xmath2 of the natural parity even multipole states in even - even nuclei @xcite .
this formula , which depends only on the mass number @xmath0 , the valence proton number @xmath3 , and the valence neutron number @xmath4 , is written as @xmath5 where @xmath6 , @xmath7 , @xmath8 , @xmath9 , @xmath10 , and @xmath11 are six model parameters to be fixed from the data for each multipole .
we have shown that this empirical formula is capable of explaining main trends of the lowest excitation energies @xmath2 of the natural parity even multipole states up to @xmath12 in even - even nuclei throughout the entire periodic table @xcite . in this work , encouraged by the apparent success of eq.([ge ] ) in describing essential trends of the lowest excitation energies @xmath2 of the natural parity even multipole states in even - even nuclei
, we plan to test whether the same empirical formula can be successfully applied to the lowest excitation energies @xmath2 of the natural parity odd multipole states .
we start with reviewing the gross behavior exposed by the data on the lowest excitation energies @xmath2 of the natural parity states in even - even nuclei .
the lowest excitation energies @xmath2 of the even multipole states including @xmath13 ( empty circles ) , @xmath14 ( solid circles ) , @xmath15 ( empty triangles ) , @xmath16 ( solid triangles ) , and @xmath12 ( empty squares ) are shown in fig.[fig-1](a ) ; while those of the odd multipole states including @xmath17 ( solid circles ) , @xmath18 ( empty triangles ) , @xmath19 ( solid triangles ) , and @xmath20 ( empty squares ) are shown in fig.[fig-1](b ) . due to a reason we give
shortly , the lowest excitation energies @xmath2 of the dipole states are shown separately by solid squares in fig.[fig-1](c ) on top of the same data shown in fig.[fig-1](b ) .
in these figures , the plotted points are connected by solid lines along the isotopic chains .
the measured excitation energies @xmath2 for the lowest @xmath13 and @xmath17 states are cited from a compilation in raman _
et al_. @xcite and kibdi _
et al_. @xcite , respectively , and those for the other multipole states are extracted from the table of isotopes , 8th edition by firestone _
et al_. @xcite .
it is quite obvious , by comparing the two graphs shown in ( a ) and ( b ) of fig.[fig-1 ] , that there are similar aspects as well as a clear distinction in the overall shape of excitation energies @xmath2 between the even multipole and odd multipole states .
the fact that the excitation energies @xmath2 are getting larger as the multipole of the state increases is still true for the odd multipole states as well as for the even multipole states .
however , the shape of the odd multipole excitation energies @xmath2 as a whole is very much different from that of the even multipole excitation energies @xmath2 because the excitation energies of odd multipole states lie significantly closer together than those of the even multipole states .
we can see also from fig.[fig-1](c ) that the lowest dipole excitation energies @xmath2 lie scattered over the region where the lowest excitation energies @xmath2 of other odd multipole states are located as shown in fig.[fig-1](b ) rather than where they are located below the lowest @xmath17 excitation energies .
this is the reason why we show the dipole excitation energy graph in addition in a separate panel ( c ) of fig.[fig-1 ] .
now , in fig.[fig-2 ] , we plot once again the excitation energies @xmath2 of exactly the same set of lowest natural parity states previously shown in fig.[fig-1 ] .
this time , however , the excitation energies @xmath2 are not taken from the data but calculated using the empirical formula , eq.([ge ] ) , with the parameter set given in tab.[tab-1 ] . the parameter values for the even multipole states presented in tab.[tab-1
] were calculated in ref.@xcite .
also , the graphs shown in fig.[fig-2](a ) are , in fact , the same as the graphs found in figs . 1 - 5 of ref.@xcite .
but , the parameter values listed in tab.[tab-1 ] for the odd multipole states are determined in this work by exactly the same method explained in ref.@xcite .
we first take the logarithmic error @xmath21 , for the @xmath22th data point , of the calculated excitation energies @xmath23 with respect to the measured excitation energies @xmath24 by @xcite @xmath25 = \log \big [ e_x^{\rm cal}(i ) \big ] - \log \big [ e_x^{\rm exp}(i ) \big].\ ] ] then the parameter values are determined by minimizing the @xmath26 value which is defined in terms of the logarithmic error by @xmath27 where @xmath28 is the number of total data points considered .
the graphs of the odd multipole states shown in fig.[fig-2](b ) and ( c ) are prepared by those parameter values . by comparing the even multipole graphs shown in fig.[fig-2](a ) with those shown in fig.[fig-1](a ) , we find that the overall shapes of the former graphs are quite similar to those of the latter graphs . furthermore , by comparing the odd multipole graphs shown in figs.[fig-2](b ) and ( c ) with those shown in figs.[fig-1](b ) and ( c ) , we also observe that the shapes of the graphs of the odd multipole excitation energies calculated using eq.([ge ] ) are quite similar to the shapes of the graphs that are determined by the data except the particular two regions which are indicated by the shaded areas in parts ( b ) and ( c ) of both figs.[fig-1 ] and [ fig-2 ] .
let us look closely into the excitation energies @xmath2 of nuclei which belong to the shaded areas in figs.[fig-1 ] and [ fig-2 ] .
for that purpose , we indicate in fig.[fig-3 ] , the measured excitation energies by filled triangles or circles while the calculated energies by empty triangles or circles .
in addition , if the measured and calculated excitation energies are relatively close , they are marked by triangles and otherwise by circles . from fig.[fig-3 ] , we confirm clearly that the excitation energies denoted by circles deviate , by quite a significant amount , from the overall trend of the excitation energies imposed by the empirical formula for all of the lowest natural parity odd multipole states .
besides , we find the following very interesting point .
the excitation energies @xmath2 marked by circles between @xmath29 belong to those nuclei whose proton number is equal to a number between 86 and 92 ; while the excitation energies @xmath2 marked by circles between @xmath30 belong to those nuclei whose neutron number is also equal to a number between 86 and 92 .
this allows us to conjecture that the nucleon numbers , equal to a number between 86 and 92 , play a certain important role in the deviation from the overall trend observed in the calculated excitation energies @xmath2 of the lowest natural parity odd multipole states .
incidentally , we have been informed later that those nuclei , whose mass number lies between @xmath30 and between @xmath29 , coincide with those in which the octupole deformation plays an important role @xcite .
the octupole deformation in nuclei is known to be produced by the long - range octupole - octupole interaction between nucleons @xcite .
the octupole correlations depend on the @xmath31 matrix elements between single particle states with @xmath32 and the energy spacing between them .
there exist particular shell structure arrangements which meet the above constraints and give rise to strong octupole correlations due to the proximity of the single particle energy levels .
three such examples include nuclei whose fermi surface lies : ( i ) between @xmath33 and @xmath34 for @xmath35 , ( ii ) between @xmath36 and @xmath37 for @xmath38 or @xmath39 , and ( iii ) between @xmath40 and @xmath41 for @xmath38 or @xmath42 .
it is very interesting to note that the nuclei between @xmath29 correspond to those whose neutron and proton orbits belong to the cases ( i ) and ( ii ) , respectively , so that the octupole - octupole interaction becomes maximum .
also , the nuclei between @xmath30 correspond again to those whose neutron and proton orbits belong to the cases ( ii ) and ( iii ) , respectively . in order to check more closely how well the empirical formula works ,
we have plotted , in each of figs.[fig-4]-[fig-8 ] , the excitation energies @xmath2 of the lowest natural parity odd multipole states including @xmath43 , @xmath44 , @xmath45 , @xmath46 , and @xmath47 , respectively , in the even - even nuclei against the mass number @xmath0 ( @xmath0-plot ) .
the upper two panels of each of these figures show the measured excitation energies while the lower two panels of each of these figures show the results for the excitation energies calculated by our empirical formula .
the plotted points of the left two panels of each of these figures are , in fact , exactly same as those shown in fig.[fig-1](b ) ( upper panels ) and in fig.[fig-2](b ) ( lower panels ) . only after the parameter values of eq.([ge ] ) were obtained employing the experimental input data from refs.@xcite , we are informed of the evaluated nuclear structure data file ( ensdf ) database which contains the most current evaluated nuclear structure information @xcite . in the right two panels of each of figs.[fig-4]-[fig-8 ] , we show the additional excitation energies extracted from the ensdf database by open circles on top of the same plotted points shown in the left two panels .
therefore , the comparison should be made between the left upper and lower panels and also between the right upper and lower panels of each of the figures . observing the graphs in the left two panels , we find that the empirical formula can reproduce the overall shape of the measured excitation energies @xmath2 of the natural parity odd multipole states in even - even nuclei reasonably well if , as mentioned earlier , we exclude the two regions indicated by the small solid horizontal bars in figs.[fig-4]-[fig-8 ] .
furthermore , it is very good to confirm , from the graphs in the right two panels , that the additional data points extracted from the ensdf database also follow the general trend predicted by our empirical formula .
the performance of our empirical formula in reproducing the lowest excitation energies @xmath2 of the natural parity odd multipole states can be inspected by drawing a scatter plot of the calculated excitation energies @xmath48 as a function of the measured energies , @xmath49 . in fig.[fig-9 ] , we show such a scatter plot of the lowest excitation energies for not only the odd multipole ( left panels ) but also the even multipole ( right panels ) natural parity states . in addition , the average @xmath50 and the dispersion @xmath51 of the logarithmic error @xmath52 defined by eq.([error ] ) for both the odd and even multipole states are shown in tab.[tab-2 ] .
as before , the scatter plot in fig.[fig-9 ] , the average @xmath50 , and the dispersion @xmath51 in tab.[tab-2 ] for the even multipole cases are quoted from ref.@xcite . by comparing the results for the odd multipole states with those for the even multipole states from fig.[fig-9 ] and tab.[tab-2 ] , we find that the performance of the empirical formula for odd multipole states is as good as , if not better , than that for the even multipole states .
moreover , we show , in the middle panels of fig.[fig-9 ] , a scatter plot of the lowest excitation energies included additionally from the ensdf database .
the recalculated @xmath26 values and the recounted total number @xmath28 of the data points after including the additional excitation energies from the ensdf database are listed inside the parentheses of tab.[tab-1 ] . the new dispersion @xmath51 of the total data points is also shown in the parentheses of tab.[tab-2 ] .
all the above results for the additional excitation energies obtained from the scatter plot , the @xmath26 values , and the dispersion @xmath51 make us to convince that the prediction power of the empirical formula is not so bad after all .
it is hard to expect that the empirical formula , like the one given by eq.([ge ] ) where @xmath2 is determined purely phenomenologically employing only the mass number @xmath0 and the valence nucleon numbers , @xmath3 and @xmath4 , can accurately reproduce the nuclear shell effects in nuclei in the vicinity of closed shells . in order to inspect this point , we redraw , in fig.[fig-10 ] , two of the scatter plots , particularly for the octupole states ( left panel ) , among the left panels of fig.[fig-9 ] , and for the quadrupole states ( right panel ) among the right panels of fig.[fig-9 ] , where the calculated excitation energies @xmath48 are plotted as a function of the measured excitation energies @xmath49 . in fig.[fig-10 ] , the plotted points are expressed by the following different symbols , according to what type of nuclei they belong to : solid circles ( both of proton and neutron shells are closed ) ; solid squares ( only one of proton and neutron shells is closed ) ; open circles ( both of proton and neutron shells are not closed ) . against our naive expectation ,
however , we only find from the scatter plot that the degree of the logarithmic dispersion from @xmath53 line is about the same regardless of shell structure of the nucleus to which a plotted point belongs .
we attribute this peculiar results to the fact that the model parameters in eq.([ge ] ) are determined simply by fitting all of the data points from nuclei throughout the entire periodic table on an equal footing .
finally , we plot the same excitation energies @xmath2 of the natural parity odd multipole states as shown in fig.[fig-4]-[fig-8 ] again in fig.[fig-11 ] but this time against the product @xmath54 ( @xmath54-plot ) instead of against the mass number @xmath0 . for years
, the @xmath54-plot has attracted much interest because a very simple pattern emerges whenever the nuclear data concerning the lowest collective state such as the lowest quadrupole excitation energy , the quadrupole transition probability , and the quadrupole deformation parameter , and so on is studied @xcite .
such a phenomenon has been called the @xmath54 scheme in the literature @xcite . in fig.[fig-11](a ) ( left panels ) , we show the measured excitation energies while in fig.[fig-11](b ) ( right panels ) , we show those calculated by the empirical formula . in these @xmath54-plots ,
the excitation energies are depicted by the following different symbols , according to which proton major shell they belong to : solid triangles ( @xmath55 ) ; empty triangles ( @xmath56 ) ; empty and solid circles ( @xmath57 ) ; and empty and solid squares ( @xmath58 ) .
particularly , the solid circles and solid squares in fig.[fig-11 ] represent the excitation energies of nuclei that do not follow the trend imposed by the empirical formula between @xmath59 and between @xmath60 , respectively , as mentioned previously in the reference to fig.[fig-3 ] . by observing fig.[fig-11](a ) ,
we find that the measured lowest excitation energies @xmath2 of the natural parity odd multipole states show almost the same simple pattern as the pattern obtained when the @xmath54-plots of the lowest excitation energies @xmath2 of the natural parity even multipole states are drawn @xcite . furthermore , by comparing the graphs in fig.[fig-11 ] ( a ) and ( b ) , we also find that the empirical formula , eq.([ge ] ) , reproduces the measured pattern of the @xmath54 plot quite closely except for the plotted points marked by solid circles and squares . and consequently , we can observe that the @xmath54 scaling fails , in the case of the natural parity odd multipole states , for nuclei where the octupole deformation prevails . in summary
, we applied the same empirical formula , which was successful in describing the essential trend of the lowest excitation energies @xmath2 of the natural parity even multipole states in even - even nuclei , to the lowest excitation energies @xmath2 of the natural parity odd multipole states .
we obtained the following three notable results .
first , in spite of the fact that the gross behavior of the odd multipole lowest excitation energies is quite different from that of the even multipole states , we observed that the empirical formula could also describe the overall trend of the measured lowest excitation energies @xmath2 of the odd multipole states up to @xmath20 .
furthermore , it turns out that the degree of the performance of the empirical formula for the odd multipole states is as good as that for the even multipole states according to the scatter plot ( fig.[fig-9 ] ) and and the dispersion @xmath51 of the logarithmic error ( tab.[tab-2 ] ) .
second , we also found that there exist two particular groups of nuclei for which the measured lowest excitation energies @xmath2 do not follow the overall trend imposed by the empirical formula as shown by fig.[fig-3 ] .
such a group of nuclei happens to be all the isotones whose neutron number is equal to a number between 86 and 92 or all the isotopes whose proton number is also equal to a number between 86 and 92 .
we emphasize that this kind of abnormality can be recognized only by a landscape description such as described in this work .
also , this abnormality may be regarded as a separate evidence of the existence of the octupole collectivity which has been known for a long time @xcite .
last , the @xmath54-plots of fig.[fig-11 ] indicate that the lowest excitation energies @xmath2 of not only the natural parity even multipole states but also the natural parity odd multipole states complies with the @xmath54 scheme .
we find it quite interesting to note that , even though the @xmath0-plots of the lowest excitation energies @xmath2 for the natural parity odd multipole states are very distinct from those for the even multipole states , the @xmath54-plots for the odd multipole states are very similar to those for the even multipole states . |
pathogens that diversify and evolve over relatively short time scales are associated with specific features of infectious disease epidemiology .
epidemics of acute respiratory infection occuring each winter in temperate climates are caused in general by influenza ( hay 2001 ) and respiratory syncytial viruses ( cane 2001 ) .
rotaviruses ( iturriza - gmara 2004 ) are the single most common cause of acute infantile gastroenteritis throughout the world .
these viruses undergo antigenic drift and both influenza a and rotaviruses undergo major antigenic shifts when a new virus is introduced into the human population through zoonotic transmission .
the dynamics of the epidemic is closely related to the antigenic structure and evolution of the viral population .
the spreading of multiple strains may be investigated by means of mathematical models , and four studies have recently uncovered the role of a ` reinfection threshold ' ( gomes _ et al _ 2005 ) in regulating pathogen diversity ( gomes _ et al _ 2002 ; boni _ et al _ 2004 ; abu - raddad & ferguson 2004 ; gkaydin _ et al _ 2005 ) . in spite of the significant differences in model formulations , assumptions , and propositions ,
there is a striking convergence at the level of the underlying mechanisms and results : levels of infection increase abruptly as a certain threshold is crossed , increasing the potential for pathogen diversity .
this may be achieved by increasing the transmissibility ( modelled by @xmath0 in this work ) , by decreasing the strength of the immunity ( modelled by increasing @xmath1 in this work ) or by increasing the rate at which novel pathogen variants are generated ( not considered in this work ) .
the concept of reinfection threshold was introduced in the context of a mean field model for vaccination impact ( gomes _ et al _ 2004 ) and has been the subject of some debate ( breban and blower 2005 ) , on the grounds that it corresponds to a change in the order of magnitude of the equilibrium density of infectives that takes place over a certain range of parameter values , rather than at a well defined threshold .
in contrast with other phenomena in epidemic models where the term threshold is widely used , the reinfection threshold does not correspond to a bifurcation , or a phase transition , occurring at a well defined critical value . in section 2
, we give a precise definition of the reinfection threshold concept , relating it to a bifurcation that takes place in the limit when the demographic renewal rate tends to zero .
thus , the reinfection threshold is a smoothed transition that occurs in systems with low birth and death rates , where a pronounced change in the systems steady state values takes place over a narrow range in parameter space , akin to a smoothed phase transition in finite physical systems around the value of the critical control parameter where a sharp transition occurs in an infinite system .
this approach reconciles the two opposing views in the literature , and we shall use the term reinfection threshold throughout the paper in this sense . for the invasion by a pathogen strain of a homogeneously mixed population where another strain circulates
, it was recently found that strain replacement is favoured when the transmissibility is below the reinfection threshold , while coexistence is favoured above ( gkaydin _ et al _ 2005 ) .
here we consider a model of contact structure in the host population to describe the underlying network for disease transmission , and investigate how this impacts on the outcome of an invasion by a new strain , that starts by infecting a very small fraction of the population .
regular lattices are simple models that account for the geographical distribution of the host population through local correlations but are not good descriptions of social contact networks as they neglect mobility patterns . by contrast ,
random graphs are simple models of complex networks that account for the mobility of the host population , through the random links or connections , but by neglecting local correlations render these models poor descriptions of real social networks .
a more realistic model of social networks was proposed recently by watts and strogatz . in 1998
they introduced small - world networks with topologies that interpolate between lattices and random graphs ( watts & strogatz 1998 ) . in these networks
a fraction of the lattice links is randomized by connecting nodes , with probability @xmath2 , with arbitrary nodes on the lattice ; the number of links is preserved by removing lattice links when random ones are established . this interpolation is non linear : for a range of @xmath2 the network exhibits small - world behaviour , where a local neighbourhood ( as in lattices ) coexists with a short average path length ( as in random graphs ) .
analysis of real networks ( dorogotsev & mendes 2002 ) reveals the existence of small - worlds in many interaction networks , including networks of social contacts . in a recent work the contact patterns for contagion in urban environments have been shown to share the essential characteristics of the watts and strogatz small - world model ( eubank _ et al _ 2004 ) . in this work
, we implement the infection dynamics on a small - world network and find that the contact structure of the host population , parametrized by @xmath2 , plays a crucial role in the final outcome of the invasion by a new strain , as departures from the homogeneously mixed regime favour strain replacement and extinction versus coexistence ( section 3 ) . within the scope of an effective mean - field model ( section 4 )
, this effect is attributed to a reduction of the effective transmissibility due to screening of infectives , resulting in a translocation of the reinfection threshold to higher transmissibility and thus decreasing the range where strain coexistence is the preferred behaviour .
this conclusion seems to contradict the established idea that localized interactions promote diversity , in the sense that the coexistence of several competing strains is favoured with respect to homogeneously mixed models . for the problem that we address here
, this idea was confirmed in a recent work ( buckee _ et al _ 2004 ) where strain competition in small - world host networks has been considered . in buckee _
et al _ , individual based stochastic simulations for a model with short term immunity , and thus high rates of supply of naive susceptibles , have shown that a localized host contact structure favours pathogen diversity , by reducing the spread of acquired immunity throughout the population .
our results indicate that when the rate of supply of susceptibles is low enough , the effect of localized interactions reported by buckee _
et al _ is superseded by that of the reduction of the effective transmissibility , that places all but the most infectious strains below the reinfection threshold .
thus , in the case of long lasting partial immunity , a predominantely local contact structure will contribute to reduce the number of coexisting competing strains in the host population .
we consider an antigenically diverse pathogen population with a two - level hierarchical structure : the population accommodates a number of strains ( two in this work ) ; each strain consists of a number of variants ( here conceptualised as many ) .
variants in the same strain differ by some average within - strain antigenic distance , @xmath3 , and different strains are separated by some antigenic distance , @xmath4 .
antigenic distance is negatively related to cross - immunity and , in some appropriate normalised measure , we may take @xmath5 .
these assumptions are consistent with the idea that rapidly evolving rna viruses experience selection as groups of antigenically close strains ( levin _ et al _ 2004 ) .
studies of influenza a viral sequence evolution ( plotkin _ et al _ 2002 ; smith _ et al _ 2004 ) provide empirical support of this modelling approach .
we consider a community of @xmath6 ( fixed ) individuals , with susceptibles and infectives in circulation .
we assume that hosts are born fully susceptible and acquire a certain degree of immunity as they are subsequently infected .
individuals are characterized by their present infection status ( healthy or infected by strain @xmath7 ) and their history of past infections ( previously infected by a set of strains ) . in this model the dynamics of a single viral strain , representing a group of cocirculating antigenically close strains , is described by a susceptible - infected - recovered ( sir ) model with partial immunity against reinfection .
abu - raddad & ferguson ( 2004 ) have proposed a different modelling approach based on the reduction of a multi - strain system to an ` equivalent ' single strain sir model , with parameters calculated to reproduce the values of the total disease prevalence and incidence of the original multi - strain model . in our model ,
the overall performance of a strain depends on a single parameter , the degree of immunity , measured by @xmath3 , for subsequent infections , and the role and meaning of the reinfection threshold ( see below for a definition ) become particularly clear . the densities ( or fractions of the host population ) are denoted by @xmath8 for susceptibles and @xmath9 for infectives .
the superscript @xmath7 of @xmath10 denotes the viral strain that currently infects that fraction of the population .
the subscripts @xmath11 describe the history of past infections ( in the case of infectives , if different from the current infecting strain ) and , for the model with two viral strains , run over the subsets of @xmath12 , that we will represent by @xmath13 , for the susceptibles . for the infectives
, @xmath11 takes the values @xmath14 , @xmath15 or @xmath16 .
the total fraction of the population infected by a given viral strain @xmath7 is @xmath17 , @xmath18 , @xmath19 , and infection of fully susceptible individuals occurs at a rate per capita ( force of infection ) given by @xmath20 .
the parameter @xmath21 is the transmissibility for strain @xmath7 .
previous infections by the same strain reduce the transmissibility by a factor @xmath1 while previous infections by a different strain reduce the transmissibility by a factor @xmath22 .
constant and equal birth / death rates , @xmath23 , and constant recovery rate @xmath24 , are also assumed .
the mean - field equations for the dynamics of the various densities of susceptibles and infectives are @xmath25 where the density of susceptibles that were previously infected with both viral strains is @xmath26
. the parameter @xmath23 is fixed at @xmath27 corresponding to a life expectancy of 80 years and the rate of recovery from infection is also fixed at @xmath28 representing an average infectious period of approximately one week .
it is well known ( anderson & may 1991 ) that a given single strain , @xmath7 , persists at endemic equilibrium if @xmath29 or , equivalently , @xmath30 where @xmath31 is the basic reproduction number of strain @xmath7 , @xmath32 the type of endemic equilibrium for the system described by ( [ mf ] ) depends on the values of the transmissibilities , @xmath33 ( or @xmath31 ) . to establish whether two strains cocirculate at equilibrium , we evaluate the eigenvalues of the jacobian at the single strain endemic equilibrium . if at least one of the eigenvalues is positive , the single strain equilibrium is unstable , and either coexistence or the other strain alone is the stable state . applying this to strain 1 , we find that this strain is expected to circulate alone when the pair @xmath34 is in the region marked 1 in figure 1(a , b ) . likewise , strain 2 is expected to circulate alone in region 2 .
the regions shaded in grey are then identified as the set of parameter values for which coexistence of both strains is stable , or coexistence regions . for comparison
, we include the dashed lines that delimit the wider coexistence regions in the absence of reinfection by the same strain ( @xmath35 ) .
the two panels correspond to systems with @xmath36 , and different values of @xmath22 : @xmath37 ( or @xmath38 ) in figure 1(a ) ; and @xmath39 ( or @xmath40 ) in figure 1(b ) .
comparing the two panels we make the expected observation that the stability of the coexisting solution is enhanced when the two strains are distantly related .
we note the existence of two small ` shoulders ' or ` kinks ' , marked a and b , on the boundaries of the coexistence regions . by numerical inspection ,
the kinks are found to be close to @xmath41 this makes intuitive sense .
a simple sir model with partial immunity ( where reinfection occurs at a rate reduced by @xmath1 ) and with a low birth and death rate @xmath23 ( as considered here ) predicts a sharp increase in the density of infected individuals at @xmath0 close to @xmath42 ( gomes _ et al _ 2004 , 2005 ) .
this region of abrupt change is related to a bifurcation at @xmath43 that occurs when @xmath44 . we shall use the term reinfection threshold for the smoothed transition that takes place at small @xmath23 , as well as for the approximate locus @xmath43 of this crossover .
focussing on strain 1 , we expect a steep increase in the infections by strain 1 as @xmath45 increases beyond @xmath46 , enhancing the competitive advantage of strain 1 relative to strain 2 and reducing the coexistence region .
two limiting cases deserve special reference : when @xmath1 is as large as @xmath22 , @xmath47 and the coexistence region collapses to the diagonal ; and when @xmath48 as in previously studied models ( may & anderson 1983 ; bremerman & thiemme 1989 ) the coexistence region is maximal . hereafter we will restrict ourselves to a symmetrical model , and will denote @xmath49 by the usual symbol for the basic reproduction number , @xmath0 .
figure 1(c , d ) shows the stability of the coexistence equilibrium as a function of @xmath0 , measured by the real parts of the seven eigenvalues of the linear approximation of system ( [ mf ] ) at equilibrium for @xmath50 and @xmath51 ( figure 1(c ) ) , @xmath40 ( figure 1(d ) ) .
all the eigenvalues have negative real parts , and all but the smallest of them change by several orders of magnitude as @xmath0 crosses the reinfection threshold region around @xmath52 .
the plot shows the modulus of the real parts of the seven eigenvalues , in logarithmic scale , versus @xmath0 , and the vertical line through @xmath52 marks the position of the reinfection threshold .
also shown by the dashed line is the total density of infectives at the equilibrium , which also increases by two orders of magnitude as @xmath0 crosses the reinfection threshold region . in both cases
we observe that as the transmissibility crosses the reinfection threshold around @xmath53 the overall stability of the coexistence equilibrium increases significantly , as the largest eigenvalues decrease steeply .
this effect is achieved by increasing either @xmath0 or @xmath1 , and the system behaviour can be characterised in terms of @xmath0 or of @xmath1 . throughout the rest of this work we fix @xmath0 and vary @xmath1 .
the behaviour of the solutions of ( [ mf ] ) for @xmath54 and @xmath55 is shown in figures 2 and 3 .
we have taken @xmath56 , @xmath36 and @xmath57 in panels a ) , b ) and c ) , respectively , corresponding to below threshold , threshold and above threshold behaviour for the fixed value of @xmath58 . by threshold we mean , as before , the window of parameter values around @xmath52 that corresponds to the smoothed transition that takes place for small birth and death rate @xmath59 , where the systems coexistence equilibrium densities and stability properties change very rapidly .
this is illustrated in the panels d ) , e ) and f ) of figures 2 and 3 , where the dashed line is at @xmath60 and the dotted line indicates the position of the reinfection threshold @xmath52 , for @xmath61 ( figure 2(d ) and 3(d ) ) , @xmath50 ( figure 2(e ) and 3(e ) ) and @xmath62 ( figure 2(f ) and 3(f ) ) .
the full line depicts the total equilibrium density of infectives with either one of the two strains as a function of @xmath0 for the ( unique ) stable steady state of the system , corresponding to endemic equilibrium where both viral strains coexist at the same steady state densities .
the numerical integration of equations ( [ mf ] ) starts with a single circulating strain ( strain 1 ) , and when the single strain steady state is reached at @xmath63 years a small fraction ( @xmath64 ) of individuals infected with the invading strain ( strain 2 ) is introduced in the system .
the two curves in panels a ) , b ) and c ) of figures 2 and 3 represent , in logarithmic scale , the infective densities for each of the strains as a function of time obtained from this numerical integration . the curve that corresponds to the density of infectives carrying strain 2 starts at @xmath63 years with density @xmath64 in all the examples shown . on relatively short
time scales of a few decades , we find that the reinfection threshold is the boundary between two different regimes .
for very similar viral strains , as depicted in figure 2 , the behaviour of the model below threshold is strain replacement , and the behaviour on and above threshold is strain coexistence . for distinct viral strains , as in figure 3
, the outcome is always strain coexistence , but the density oscillations are negligible above threshold , and very pronounced below threshold . in order to extract epidemiologically significant predictions from the mean field model we have to take into account that the population is discrete , and that densities below a certain lower bound , that depends on the population size , are effectively zero .
if we take this effect into account and set the densities to zero when they fall below @xmath65 , we obtain extinction of both strains in the case of figure 3(a ) . for longer time scales
, however , the model also predicts coexistence for the case of similar strains below threshold ( figure 2(a ) ) since there are no attractors in the single strain invariant subspaces when the two strains have the same transmissibility .
all the solutions that were analysed tend eventually to the coexistence equilibrium , but the transients for similar strains below threshold last for several hundred years , during which the density of infectives carrying strain 1 reaches much smaller values than the cut - off of @xmath65 .
this behaviour and the role of the reinfection threshold can be interpreted in terms of the eigenvalues of the coexistence endemic equilibrium represented in figure 1(c , d ) . as @xmath0 increases across the threshold ,
the real parts of the eigenvalues decrease steeply , and the imaginary parts vanish . as a result
the oscillatory component of the system plays an important role below the reinfection threshold and is absent above .
this is the basic ingredient of the different invasion dynamics depicted in figures 2 and 3 .
the difference in below threshold behaviour between similar and dissimilar strains , strain replacement in the former case and total extinction in the latter may be understood intuitively in terms of the reduced cross - immunity of dissimilar strains . in this case , the small fraction of infectives carrying strain 2 finds a highly susceptible population and generates a large epidemic outbreak , during which strain 1 goes extinct because of the lack of suceptibles .
strain 2 quickly exhausts its pool of susceptibles through first infections , and then dies out too as @xmath66 .
then , for realistic population sizes , the predictions of the mean - field model are the following
. for similar viruses , strain replacement ( drift ) occurs below threshold , and coexistence on and above threshold . for dissimilar viruses
, global extinction occurs below threshold while coexistence occurs on and above threshold .
it is known ( bayley 1975 ) that fluctuations due to stochastic effects in discrete populations will also favour strain extinction and replacement with respect to deterministic model predictions . a homogeneously mixed stochastic version of model ( [ mf ] ) was considered by gkaydin _
et al _ ( 2005 ) for population sizes between @xmath67 and @xmath68 , and , as expected , stochasticity was found to favour strain replacement and extinction both for similar and dissimilar strains .
this effect may be , however , greatly enhanced in more realistic descriptions of the host population where the homogeneously mixed assumption is relaxed .
in the presence of random long - range links such as those considered in small - world networks the endemic and epidemic thresholds of susceptible - infective - susceptible ( sis ) and sir models may be mapped on to _ mean - field _ site and bond percolation transitions ( grassberger 1983 , dammer & hinrichsen 2003 , hastings 2003 ) . in recent works ,
the network topology has been considered in the calculation of endemic and epidemic thresholds of sis and sir models ( may & lloyd 2001 , moore & newmann 2000a , 2000b , pastor - santorras & vespigniani 2001a ) and the results revealed a strong dependence of the threshold values on the network size and structure .
the contact network topology has also been shown to play an important role in the stationary properties of the endemic state ( pastor - santorras & vespigniani 2001b ) , the short term dynamics of epidemic bursts ( keeling 1999 , kleczkowski & grenfell 1999 , rhodes & anderson 1996 ) , the long term dynamics of childhood diseases ( verdasca _ et al _ 2005 ) and in the estimation of disease parameters from epidemiological data ( meyers _ et al _ 2005 ) .
the structure of the network of contacts of the host population is also expected to play a role in the invasion dynamics described in the previous section .
this will affect the pattern of strain replacement and as a consequence the evolution of competing multi - strain pathogens . in the following we analyse quantitatively the effects of a network of social contacts with small - world topology on the simple two strain model ( [ mf ] ) .
we implemented a discrete version of the model ( [ mf ] ) on a cellular automaton ( square lattice ) with @xmath69 sites and small - world interaction rules ( see the appendix for a description of the algorithm ) .
we started by simulating the behaviour of the system for a single strain model , whose mean field equations are simply @xmath70 we found characteristic medium and long - term dynamics related , in a quantitative fashion , to the structure of the network of contacts .
in particular , as @xmath2 decreases , the increase in spatial correlations ( i ) decreases the effective transmissibility through the screening of infectives and susceptibles , which in turn increases the value of the transmissibility at the endemic and reinfection thresholds .
in addition , the spatial correlations ( ii ) enhance the stochastic fluctuations with respect to the homogeneously mixed stochastic model .
this effect is particularly strong at low @xmath2 , where the relative fluctuations are largest and where as a consequence ( iii ) the dependence of the steady state densities on the effective transmissibility predicted by the mean - field equations breaks down .
analogous effects , including departures from the mean - field behaviour at the endemic threshold ( persistence transition ) of a sir model with non - zero birth rate have been reported recently ( verdasca _ et al _ 2005 ) . in figure 4
we plot the effective transmissibility , @xmath71 ( average density of new infectives per time step divided by the value of @xmath72 for the instantaneous densities ) , at fixed @xmath73 and @xmath74 , well above the reinfection threshold of the mean - field model , as a function of the small - world parameter , @xmath2 .
the variation of the effective transmissibility with @xmath2 is due to the clustering of infectives and susceptibles .
this is a well known screening effect that results from the local structure ( correlations ) , and has to be taken into account in fittings to effective mean - field models . in this framework
, the reduction of the effective transmissibility represents an increase of the value of @xmath0 at the reinfection threshold .
the clustering , and spatial correlations in general , have also drastic consequences on the amplitude and nature of the stochastic fluctuations .
enhanced fluctuations will ultimately lead to stochastic extinction as @xmath2 decreases , but before extinction occurs , the fluctuations lead to the appearance of a regime dominated by local structure and correlations , where the mean - field relations among the equilibrium densities break down .
this is illustrated in figure 5 where we plot the effective transmission rate ( dots ) as a function of the equilibrium density of infectives ( model parameters as in figure 4 ) . in the same figure we plot the equilibrium value of the density of infectives ( full line ) predicted by the mean - field equations ( [ mf1 ] ) for a value of the transmissibility equal to the effective transmissibility obtained from the simulations .
when @xmath75 , away from the network endemic threshold at @xmath76 , the mean - field equilibrium density for the effective transmissibility is in excellent agreement with the simulated equilibrium density . in the region of small @xmath2 ( @xmath77 ) ) , however , the simulated equilibrium densities differ significantly from those calculated using the effective mean - field theory .
a departure from the mean - field dependence of the effective transmissibility on the steady state density of infectives implies that the description of the system is no longer possible in terms of effective mean - field theories based on density dynamics , by contrast with the second regime where the contact structure can be implicitely taken into account by effective mean - field models ( section 4 ) .
we then focussed on the analysis of the invasion dynamics .
we performed individual based simulations starting from an initial condition where the system is close to the steady state for a single resident strain and introduced a small fraction of individuals ( @xmath64 ) infected by the invading strain .
the algorithm is a natural extension of the single strain stochastic algorithm for sir dynamics with reinfection on a small - world network of contacts , and a detailed description is given in the appendix .
we found that as local effects become important ( low @xmath2 ) , strain replacement ( and also total extinction ) are favoured with respect to the homogeneously mixed regime ( @xmath78 ) .
these results are summarized in figure 6 where we plot @xmath1 at crossover between the different regimes , coexistence , replacement and extinction , versus the small - world parameter , @xmath2 , for systems where the competing viral strains are similar , @xmath79 ( figure 6(a ) ) and dissimilar , @xmath80 ( figure 6(b ) .
the full line is the boundary between the strain coexistence and the strain replacement regimes , and the dashed line is the boundary between strain replacement and total extinction .
the final outcome of an invasion is determined by carrying out a series of simulations , for different values of @xmath2 and @xmath1 , and keping @xmath0 fixed at @xmath81 . for each @xmath82 , twenty ( forty in some cases )
invasion simulations are performed for a period of 13.7 years , with invasion at @xmath83years . the fraction of simulations where both the circulating and the invading strain prevail is plotted as a function of @xmath1 , at fixed @xmath2 .
this fraction changes rapidly from one to zero as @xmath3 decreases across a small interval , and the boundary of the coexistence regime ( the full line in figure 6(a , b ) is determined as the point where it takes the value @xmath84 . a similar analysis yields the boundary that separates the replacement from the total extinction regime ( the dashed line in figure 6(a , b ) .
the results show that as @xmath2 decreases the value of @xmath1 at these two boundaries increases exponentially from its value in the homogeneously mixed system , @xmath78 . as a consequence ,
the range of parameters in the coexistence regime decreases drastically as @xmath2 decreases , supporting the claim that the structure of the small - world network of contacts hinders , rather than favours , pathogen diversity .
this contrasts with the behaviour reported by buckee _
et al _ ( 2004 ) for a model of strain evolution including short term host immunity and cross - immunity on a ( static ) small - world network of contacts , where coexistence of competing strains was found to be favoured with respect to the homogeneously mixed model .
the different behaviour we report is due to the combined effect of small - world network structure and a low rate of supply of naive susceptibles , and may be understood in terms of the analysis of the single strain stochastic model discussed previously , together with the behaviour of the two - strain mean - field model which , for low @xmath59 , predicts strain coexistence as the outcome of an invasion only when the transmissibility is high enough
. recall that in the mean - field model of section 2 ( system ( [ mf ] ) ) when a second strain is introduced in a population with a resident virus , drift occurs below the reinfection threshold for antigenically similar strains ( small @xmath85 ) . when the resident and the invading strains are antigenically distant ( large @xmath85 ) the final outcome is coexistence at and above the reinfection threshold and global extinction below threshold .
the results of the individual based simulations show that the enhancement of the stochastic fluctuations due to spatial correlations favours replacement where we would otherwise have coexistence , and global extinction where we would otherwise have replacement .
in particular , replacement ( both drift and shift , replacement of dissimilar strains ) becomes a typical outcome at the reinfection threshold , instead of coexistence as in the mean - field model .
this may be viewed as a finite size stochastic effect similar to the effect reported in gkaydin _
et al _ 2005 for the homogeneously mixed stochastic model , or to what we have also found here for @xmath78 , but much more pronounced . as we noted previously the amplitude of the stochastic fluctuations at small values of @xmath2
is greatly enhanced with respect to the amplitude of the stochastic fluctuations at @xmath78 , due to coherent fluctuations of host population clusters .
however , the main effect of spatial correlations that accounts for the results of figure 6 is the screening of infectives that leads to the reduction of the effective transmissibility as shown in figure 4 .
as @xmath2 decreases , this effect brings a system above the reinfection threshold closer to or even below the reinfection threshold , also favouring replacement for small @xmath86 and either replacement or extinction for larger @xmath86 .
while the effects of the coherent fluctuations of clusters are important only in a small range of @xmath2 above the single strain endemic transition , screening of infectives occurs over the whole range of @xmath2 with the corresponding translocation of the reinfection threshold to larger values of @xmath0 .
in particular , for a large range of values of @xmath2 we obtain drift / shift , instead of coexistence as predicted by the well mixed model for the same disease parameters .
many of these effects may be captured by effective mean - field models as described in the next section .
in section 3 we have seen that , for single strain dynamics , the mean field equilibrium incidence given by ( [ mf1 ] ) for the effective transmissibility is in excellent agreement with the results of the simulations for values of the small world parameter @xmath2 in an interval @xmath87 $ ] , while departure from the mean field relations occurs close to the endemic transition threshold at @xmath77 ( see figure 5 ) . in this section
we show that , for the full system and @xmath2 in the range @xmath87 $ ] , the simulated time series are well approximated by the solutions of an effective mean - field model of the form ( [ mf ] ) . for the construction of the effective model
we assume that the transmissibility is of the form @xmath88 , where @xmath89 is a function , unknown a priori , that represents the screening of infectives .
the screening function @xmath89 is obtained from a plot @xmath90 as in figure 4 , where the screeening effect is quantified for the single strain dynamics .
the relevant range of @xmath2 for this fit is @xmath87 $ ] , since the breakdown of the mean field equilibrium relations implies that our ansatz for the form of the infection rate is no longer valid . even in this range
it is not obvious that the ansatz will yield an effective mean - field model capable of describing the full dynamical behaviour of the simulations when a resident and an invading strain interact . in figure 7(a ) , black line , we show the results of stochastic invasion simulations as described in section 3 and in the appendix , for @xmath91 ( other parameters as in figure 2.b ) ) , together with the numerical integrations , grey line , of the coupled strain effective mean - field model , given by ( [ mf ] ) with @xmath92 , for @xmath93 . also shown in figure 7(b ) are , for comparison , the solutions of the standard mean - field model , where @xmath94 ( grey line ) , plotted together with the stochastic simulations of figure 7(a ) ( black line ) . in all cases
we use the same initial conditions and the same invasion conditions .
while the standard mean - field model disagrees quantitatively , and even qualitatively , with the results of the simulations , the effective mean - field model describes accurately the dynamical behaviour of the system , except for the stochastic fluctuations about the average densities .
systematic calculations show that the effective mean - field model performs rather well against simulations , everywhere in the region where the mean - field relations apply . in this regime
, the effect of the contact structure of the host population is accurately described through the screening function @xmath89 , obtained from the steady state averages of single strain simulations . in the range
@xmath95 $ ] of @xmath2 above the endemic threshold where we have found significant departures from the mean - field description , one may think that a more general ansatz for the form of the force of infection could lead to a modified effective mean - field model , albeit a more complicated one , that would fit the simulation results .
however , it is easy to check that the parametric plot of @xmath96 vs @xmath97 , where @xmath97 is the steady state average infective density , follows the mean field relation of the model with linear force of infection whatever the functional form assumed for @xmath98 . indeed , for any function @xmath98 , and rate of infection of the form @xmath99 @xmath100 measured from the simulations is @xmath101 , and the curve @xmath102 is the same as the standard mean field curve for @xmath103 .
this means that even in the scope of a more general model with a nonlinear force of infection the mean field curve for @xmath96 vs @xmath97 , the full line in figure 5 , does not fit the simulations results for @xmath2 in the range @xmath95 $ ] .
we conclude that the effect of the breakdown of the mean - field relations reported in section 3 is an indication of a new regime where spatial correlations are too important for the contact rate between individuals of different classes to be described by the product of the corresponding densities .
the construction of effective models in this regime requires the use of pair approximations that take into account the spatial correlations ( de aguiar _ et al _ 2003 ) and will be the subject of future work .
we performed stochastic simulations of individual based models on small - world networks to represent host populations where one or two viral strains , each representing a group of antigenically close variants , may be present . in the case of a single viral strain , the node dynamics corresponds to a sir model with reinfection at a reduced rate due to acquired immunity . in the case of two competing strains ,
the rate of infection by a distinct strain is also reduced due to cross immunity , which is always weaker than strain specific immunity .
both the single and the double strain models include crucial ingredients required by realistic modeling of the host population , namely stochasticity , a discrete finite population , and spatial structure given by a plausible contact network .
we analised both the reinfection dynamics of the single strain model and the invasion dynamics of the two strains model , taking initial conditions that correspond to the presence of a small number of individuals infected with a pathogen strain in a population where another pathogen strain circulates . for single - strain dynamics
, we found that the major effect of spatial correlations is a decrease in the effective transmissibility through the screening of infectives and susceptibles which in turn increases the value of the transmissibility at the endemic and at the reinfection thresholds .
in addition , spatial correlations enhance the amplitude of the stochastic fluctuations with respect to the homogeneously mixed stochastic model .
this effect is particularly strong at low @xmath2 , where the relative fluctuations are largest and where the mean - field dependence of the steady state densities on the effective transmissibility breaks down .
indeed , we have found a regime at small @xmath2 , where spatial correlations dominate and the mean - field relations break down , as well as a second , wider , regime for larger values of @xmath2 where effective mean - field models are capable of describing the essential effects of the spatial correlations through a reduced effective transmissibility .
for the two - strain model , we found that the host contact structure significantly affects the outcome of an attempted invasion by another strain of a population with an endemic resident strain , and by contrast with standard expectations we observed that spatial structuring reduces the potential for pathogen diversity .
the simulation results show that the structure of the network of contacts favours strain replacement ( drift / shift ) or global extinction versus coexistence as the outcome of an invasion , with respect to the mean - field and stochastic homogeneously mixed populations .
in particular , we find that as the small - world parameter , @xmath2 , decreases at fixed @xmath0 , the value of the strain specific immunity parameter @xmath3 above which coexistence is the typical outcome increases exponentially from its value for the homogeneously mixed system , @xmath78 , supporting the claim that local correlations strongly reduce pathogen diversity .
this conclusion is in apparent contradiction with the established idea that localized interactions promote diversity , as confirmed by stochastic simulations for a model with high rates of supply of naive susceptibles . in that model
the effect of the host contact structure favours pathogen diversity , by suppressing the uniform spread of acquired immunity throughout the population .
by contrast , in our model characterised by partial permanent immunity and low rates of supply of naive susceptibles ( through births ) the former effect is superseded by the screening effect .
the latter is also due to the clustering of the host population , and results in a reduction of the effective transmissibility , changing the reinfection threshold to higher levels of @xmath0 , and therefore reducing the range where coexistence is the preferred behaviour .
the results of our simulations strongly support the conclusion that in systems with a reinfection threshold , due to a low rate of supply of naive susceptibles , the major effect of the host population spatial structuring is the effective reduction of pathogen diversity . the immunity profile ( ferguson & al .
2003 ) and the duration of infection ( gog & grenfell 2002 ) have recently been shown to play an important role in shaping the patterns of pathogen evolution in multi - strain models with mutation generated antigenic diversity . investigating the additional effect of host population contact structure topology in these models will be the subject of future work .
financial support from the portuguese foundation for science and technology ( fct ) under contracts pocti / esp/44511/2002 , pocti / isfl/2/618 and pocti / mat/47510/2002 , and from the european commission under grant mext - ct-2004 - 14338 , is gratefully acknowledged .
the authors also acknowledge the contributions of j. p torres and m. simes to test and to improve the code used in the simulations .
99 abu - raddad , l j and ferguson , n f ( 2004 ) the impact of cross - immunity , mutation and stochastic extinction on pathogen diversity , _ proc .
_ b * 271 * 2431 - 2438 .
+ de aguiar , m a m , rauch , e m , bar - yam , y ( 2003 ) mean field approximation to a host - pathogen model , _ phys
e _ * 67 * 047102 .
+ anderson , r a and may , r m ( 1991 ) _ infectious diseases of humans _ , oxford u. p. , oxford .
+ bailey , n t j ( 1975 ) _ the mathematical theory of infectious diseases _ , 2nd edition , charles griffin & co , london .
+ boni , m f , gog , j r , andreasen , v and christiansen , f b ( 2004 ) influenza drift and epidemic size : the race between generating and escaping immunity , _ theor .
_ * 65 * 179 - 191 .
+ bremermann , h and thieme , h r ( 1989 ) a competitive - exclusion principle for pathogen virulence , _ j. math
. biol . _ * 27 * 179 - 190 .
+ buckee , c o f , koelle , k , mustard m j , and gupta , s ( 2004 ) the effects of host contact network structure on pathogen diversity and strain structure , _ proceedings of the national academy of science _ * 101
* 10839 - 10844 . + cane , p c ( 2001 ) molecular epidemiology of respiratory syncytial virus , _ rev . med .
virol . _ * 11 * 103 - 116 .
+ dammer , s m and hinrichsen , h ( 2003 ) epidemic spreading with immunization and mutations , _ phys .
e _ * 68 * 016114 .
+ dorogotsev , s n and mendes , j f ( 2003 ) _ evolution of networks _
, oxford : oxford university press .
+ grassberger , p ( 1983 ) on the critical behaviour of the general epidemic process and dynamical percolation _ math .
biosci . _ * 63 * 157 - 172 .
+ eubank , s , guclu h , anil kumar v s , marathe m v , srinivasan a , torockzcal z and wang n ( 2004 ) modelling disease outbreaks in realistic urban social networks , _ nature _ * 429 * 180 - 184
. + ferguson , n m , galvani , a p and bush , r b ( 2003 ) ecological and immunological determinants of influenza evolution , _ nature _ * 422 * 428 - 433 . + gog , j r and grenfell , b t ( 2002 ) dynamics and selection of many - strain pathogens , _ proc .
natl acad .
usa _ * 99 * 17209 - 17214 .
+ gkaydin , d , oliveira - martins , j b , gordo , i and gomes , m g m ( 2005 ) the reinfection threshold regulates pathogen diversity : the unusual patterns of influenza a evolution ( submitted ) .
+ gomes , m g m , medley , g f and nokes , d j ( 2002 ) on the determinants of population structure in antigenically diverse pathogens _ proc .
_ b * 269 * 227 - 233 .
+ gomes , m g m , white , l j and medley , g f ( 2004 ) infection , reinfection , and vaccination under suboptimal immune protection : epidemiological perspectives .
_ j. theor .
biol . _ * 228 * 539 - 549 .
+ gomes , m g m , white , l j and medley , g f ( 2005 ) the reinfection threshold , _
( in press ) .
+ hastings , m b ( 2003 ) mean - field and anomalous behaviour on a small world network , _ phys .
lett . _ * 91 * 098701 .
+ hay , a j , gregory , v , douglas , a r and lin , y p ( 2001 ) the evolution of human influenza viruses , _ philos .
_ b * 356 * 1861 - 1870 . + iturriza - gmara , m , kang , g and gray , j ( 2004 ) rotavirus genotyping : keeping up with an evolving population of human rotaviruses , _
* 31 * 259 - 265 .
+ keeling , m j ( 1999 ) the effects of local spatial structure on epidemiological invasions _ proc .
lond . b _ * 266 * 859 - 867 .
+ kleczkowski , a and grenfell , b t ( 1999 ) mean field type of equations for spread of epidemics : the small world model , _ physica a _ * 274 * 355 - 360
. + levin , s a , dushoff , j and plotkin , j b ( 2004 ) evolution and persistence of influenza a and other diseases , _ math .
biosci . _ * 188 * 17 - 28 .
+ may , r m and anderson , r m ( 1983 ) epidemiology and genetics in the coevolution of parasites and hosts , _ proc .
_ b * 219 * 281 - 313 .
+ may , r m and lloyd , a l ( 2001 ) infection dynamics on scale free networks , _ phys .
e _ * 64 * 066112 .
+ meyers , l a , pourbohloul , b p , newman m e j , skowronski d m , brunham r c ( 2005 ) network theory and sars : predicting outbreak diversity , _ j. theor .
_ * 232 * 71 - 81 .
+ moore , c and newmann , m e j ( 2000 ) epidemics and percolation in small world networks , _ phys .
e _ * 61 * 5678 - 5682 .
+ moore , c and newmann , m e j ( 2000 ) exact solution of site and bond percolation on small world networks , _ phys .
e _ * 62 * 7059 - 7064 .
+ pastor - santorras , r and vespigniani , a ( 2001 ) epidemic spreading in scale free networks , _ phys .
* 86 * 3200 - 3203 .
+ pastor - santorras , r and vespigniani , a ( 2001 ) epidemic dynamics and endemic states in complex networks , _ phys .
e _ * 63 * 066117 .
+ plotkin , j b , dushoff , j and levin , s a ( 2002 ) hemagglutinin sequence clusters and the antigenic evolution of influenza a virus , _ proceedings of the national academy of science _ * 99 * 6263 - 6268 .
+ rhodes , c j and anderson , r m ( 1996 ) persistence and dynamics in lattice models of epidemic spread , _ j. theor .
* 180 * 125 - 133 .
+ smith , d j , lapedes , a s , de jong , j c , bestebroer , t m , rimmelzwaan , g f , osterhaus , a d and fouchier , r a ( 2004 ) mapping the antigenic and genetic evolution of influenza virus , _ science _ * 305 * 371 - 376 .
+ verdasca j , telo da gama , m m , nunes , a , bernardino , n r , pacheco , j m and gomes , m c ( 2005 ) recurrent epidemics in small world networks , _ j. theor .
* 233 * 553 - 561
. + watts , d j and strogatz , s h ( 1998 ) it s a small world , _ nature _ * 392 * 440 - 442 .
a community of @xmath6 ( fixed ) individuals comprises , at time @xmath104 , susceptibles and infectives in circulation .
hosts are born fully susceptible and acquire a certain degree of immunity as they are subsequently infected .
we consider susceptibles and infectives that were previously infected by strain @xmath10 , @xmath105 and @xmath106 .
the indices denote previous and current infections as in the text .
we consider a cellular automaton ( ca ) on a square lattice of size @xmath107 with periodic boundary conditions .
the ( random ) variables at each site may take one of eight values : @xmath108 , @xmath109 , @xmath110 , @xmath111 , @xmath112 , @xmath113 , @xmath114 or @xmath115 .
the lattice is full .
we account for local interactions / connections with @xmath7 neighbouring sites , with @xmath116 , and long - range interactions / connections , with a small - world probability , @xmath2 .
the transmissibility @xmath117 is the sum of the local and long - range rates of transmission .
first infection , within - strain and between - strains reinfection , recovery , birth and death occur stochastically , with fixed rates ( @xmath117 , @xmath118 , @xmath119 , @xmath24 , @xmath23 , @xmath120 .
the recovery time ( @xmath121 ) sets the time scale . at each monte carlo ( time ) step , @xmath6
random site updates are performed following a standard algorithm .
the type of event , long or short - range infection , within and between - strains reinfections by strains 1 and 2 , recovery , birth and death , is chosen with the appropriate frequency ( @xmath122 , @xmath123 , @xmath124 , @xmath125 , @xmath126 , @xmath127 , @xmath24 , @xmath23 , @xmath23 ) and then proceeds as follows .
long(short)-range first infection by 1 ( 2 ) .
a site is chosen at random ; if the site is occupied by an @xmath128 or an @xmath129 other than @xmath108 no action is taken .
if the site is occupied by an @xmath108 , one of the other lattice sites ( or of its @xmath7 neighbours for short - range infection ) is chosen at random ; @xmath108 is infected by 1 ( 2 ) iff that site is an @xmath130 ( @xmath131 ) .
long(short)-range between strain reinfection by 1 ( 2 ) .
a site is chosen at random ; if the site is occupied by an @xmath128 or an @xmath129 other than @xmath110 ( @xmath109 ) no action is taken .
if the site is occupied by an @xmath110 ( @xmath109 ) , one of the other lattice sites ( or of its @xmath7 neighbours for short - range infection ) is chosen at random ; @xmath110 ( @xmath109 ) is reinfected by 1 ( 2 ) iff that site is an @xmath130 ( @xmath131 ) .
long(short)-range within strain reinfection by 1 ( 2 ) .
a site is chosen at random ; if the site is occupied by an @xmath128 or an @xmath129 other than @xmath109 ( @xmath110 ) or @xmath115 no action is taken .
if the site is occupied by an @xmath109 ( @xmath110 ) or @xmath115 one of the other lattice sites ( or its @xmath7 neighbours for short - range infection ) is chosen at random ; @xmath109 ( @xmath110 ) or @xmath115 is reinfected by 1 ( 2 ) iff that site is an @xmath130 ( @xmath131 ) .
* figure 1 * stability analysis .
( a , b ) conditions for strain coexistence or competitive exclusion as described by the two strain mean - field model .
the panels correspond to models with @xmath133 and different values of @xmath22 : ( a ) @xmath37 ; ( b ) @xmath39 .
the regions shaded in grey are the set of parameter values for which coexistence of both strains is stable , or coexistence regions .
the dashed lines delimit the wider coexistence regions in the absence of reinfection by the same strain ( @xmath35 ) . in region 1
2 ) , the single strain endemic equilibrium for strain 1 ( resp .
2 ) is stable .
( c , d ) modulus of the real parts of the seven eigenvalues of the coexistence equilibrium along the diagonal of ( a , b ) respectively , in logarithmic scale .
also shown is the total density of infectives at equilibrium ( dashed line ) , and the position of the reinfection threshold ( dotted line ) .
* figure 2 * mean - field new infectives densities for the two strain model of equations ( 1 - 4 ) with @xmath54 .
we have taken @xmath134 , @xmath36 and @xmath57 in panels a ) , b ) and c ) , respectively , corresponding to below threshold , threshold and above threshold behaviour for the value of @xmath58 .
this is illustrated in the diagrams d ) , e ) and f ) where the dashed line is at @xmath60 and the dotted line indicates the position of the reinfection threshold @xmath135 .
the full line depicts the equilibrium total density of infectives , @xmath136 , as a function of @xmath0 for the stable steady state of the system , the endemic equilibrium where both viral strains coexist .
* figure 3 * mean - field new infectives densities for @xmath55 . as in figure 2
we take @xmath134 , @xmath36 and @xmath57 in panels a ) , b ) and c ) , respectively , corresponding to below threshold , threshold and above threshold behaviour for @xmath58 , as illustrated in the diagrams of panels d ) , e ) and f ) . *
figure 4 * effective transmissibility vs the small - world parameter , @xmath2 .
the effective transmissibility , @xmath71 , is calculated as the average density of new infectives per time step divided by @xmath137 where @xmath138 and @xmath10 are the instantaneous densities of susceptibles and infectives .
the drastic reduction in @xmath71 is due to the clustering of infectives and susceptibles as @xmath2 decreases and the spatial correlations increase .
the model is the single strain reinfection model ( [ mf1 ] ) and the parameters are @xmath73 and @xmath139 .
the number of nodes is @xmath132 .
* figure 5 * effective transmissibility vs the equilibrium infective density , @xmath97 ( dots ) , over the whole range of the small - world parameter , @xmath2 ( dots ) .
notice the logarithmic scale in the @xmath97 axis .
the full line is the @xmath140 curve calculated from the mean - field single strain reinfection model ( [ mf1 ] ) .
the parameters are as in figure 4 .
the results show the existence of two regimes : an effective mean - field regime for @xmath141 , where the relations between the equilibrium densities are given by the mean - field equations for the screened value of the transmissibility , and a fluctuation dominated regime for @xmath77 , where the mean - field relations break down .
* figure 6 * coexistence versus replacement ( full line ) and replacement versus total extinction ( dashed line ) crossovers in @xmath82 space .
the reinfection parameter , @xmath1 , is plotted as a function of the small - world parameter , @xmath2 , at the crossover separating the dynamical regimes of coexistence of two competing viral strains ( above the full line ) , of invasion prevalence ( between the dashed and the full lines ) and of total extinction ( below the dashed line ) .
see the text for details .
panel a ) corresponds to similar strains , @xmath142 , and plot b ) to dissimilar strains , @xmath143 .
@xmath58 and @xmath6 as in other simulations .
* figure 7 * mean field and individual based description of invasion dynamics .
the same initial conditions and the same invasion conditions are used in all cases .
( a ) mean - field new infectives densities ( grey line ) for the model of equations ( [ mf ] ) with @xmath92 , @xmath93 , fitted from the simulations of the single strain model for @xmath91 ( other parameters as in figure 2(b ) ) , and time series ( black line ) of the stochastic model for @xmath91 ( other parameters as in figure 2(b ) ) .
( b ) we also show , for comparison , the solutions of the standard mean - field model ( grey line ) , where @xmath144 , @xmath93 , are not corrected to take into account the screening
. the black line corresponds to the same data as in ( a ) . | we investigate the dynamics of a simple epidemiological model for the invasion by a pathogen strain of a population where another strain circulates .
we assume that reinfection by the same strain is possible but occurs at a reduced rate due to acquired immunity .
the rate of reinfection by a distinct strain is also reduced due to cross - immunity .
individual based simulations of this model on a ` small - world ' network show that the host contact network structure significantly affects the outcome of such an invasion , and as a consequence will affect the patterns of pathogen evolution .
in particular , host populations interacting through a small - world network of contacts support lower prevalence of infection than well - mixed populations , and the region in parameter space for which an invading strain can become endemic and coexist with the circulating strain is smaller , reducing the potential to accommodate pathogen diversity .
we discuss the underlying mechanisms for the reported effects , and we propose an effective mean - field model to account for the contact structure of the host population in small - world networks .
keywords : pathogen diversity , reinfection threshold , spatial structure , complex networks |
x - ray binaries harbour compact objects and their x - ray luminosity is believed to be generated due to an accretion disk formed by material accreted from a companion star .
the x - ray emission is known to vary on a wide range of timescales .
it was after the launch of the _ rossi x - ray timing explorer _
( rxte ) satellite , that millisecond variability was detected in several sources ( for reviews see for e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
the x - ray variability is typically quantified in the frequency domain by computing the power density spectra .
the power density spectrum sometimes shows peaked features which are termed as quasi - periodic oscillations ( qpo ) in some neutron star low - mass x - ray binaries ( lmxbs ) , there are qpos observed in the frequency range of 400 to 1300 hz , and these are termed as kilohertz qpos . often two simultaneous khz qpos are observed in a system , with the higher - frequency one being called the ` upper khz qpo ' while the lower - frequency one is called the ` lower khz qpo ' . in general , the quality factor ,
@xmath2 , which is the ratio of the qpo frequency to the full width at half maximum , of the lower khz qpo is significantly greater than the upper one @xcite .
there have been several models invoked to explain this twin khz qpo phenomena .
a list of different models with their merits and demerits has been presented by @xcite . here
we briefly mention some of them . in the sonic - point model
@xcite with modifications @xcite , the frequency of the upper khz qpo is interpreted as the keplerian one at the sonic radius , while the frequency of lower khz qpo is generated by the modulation of the upper qpo with the neutron star spin frequency . in this interpretation , the frequency separation between the upper and lower khz qpo should be a constant , but almost all sources show significant variation in the separation .
for example , for the lmxb 4u 1608 - 52 a linear relation between the upper and lower khz qpo has been reported @xcite . the relativistic precession model @xcite identifies the upper khz qpo frequency as a keplerian one with the lower khz qpo generated due to modulation of the keplerian frequency with the relativistic periastron precession frequency . in the two - oscillator model @xcite
, the lower khz qpo is due to a blob moving in a keplerian orbit , and the upper khz qpo is generated by the influence of the coriolis force on the blob when it enters the neutron star s rotating magnetosphere .
later , @xcite refined this model by connecting the qpo to a rayleigh - taylor gravity wave .
@xcite have proposed that the accretion disk around a rapidly rotating compact object may be misaligned beyond a transition radius due to the bardeen - petterson effect and this gives rise to the qpo phenomenon
. there have also been attempts to understand these qpos as global oscillation in the disk due to hydrodynamic or magneto - hydrodynamic ( mhd ) instabilities or waves . in a series of papers , ( e.g. * ? ? ?
* ; * ? ? ?
* ) the qpo is explained to be generated by a resonant excitation caused by the non - linear interactions between the disk oscillation modes .
the phenomena has been attributed to alfven wave oscillations @xcite or to standing modes of mhd waves @xcite .
the rossby wave instability in the inner disk has also been invoked @xcite . @xcite
have tried to explain the qpos in terms of a resonant coupling of mhd oscillation modes with the neutron star spin . @xcite
have performed mhd simulations to understand the phenomena .
apart from theoretical consistences , the validity of these different models can in principle be tested with observations .
a model should explain the observed relationships between the twin khz qpo frequencies and any other lower frequency qpos ( e.g. * ? ? ?
* ; * ? ? ?
it may also need to explain the strength of the khz qpos ( e.g. * ? ? ?
* ; * ? ? ?
* ) or the condition for the appearance of two khz qpos ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
unfortunately , there is no consensus that any of the models is the most favourable one .
this is primarily because none of the models have a clear predictive and distinguishing signature which can be tested against these observations .
it is also not clear how does a dynamical oscillation proposed by these different models , couples to the radiative process that produces the x - ray emission .
relatively fewer studies have been undertaken to identify the radiative mechanism that is responsible for the khz qpos .
on the other hand , the time averaged spectra of lmxbs have been well studied by several missions .
the primary component has been indisputably identified to be thermal comptonization which typically dominates the spectrum especially at energies @xmath3 kev .
there is a secondary soft competent which could be a multi - coloured disk emission ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) or a black body from the boundary layer between the disk and the star ( e.g. * ? ? ?
* ; * ? ? ?
the spectra show significant and systematic long term variation which is characterised by shapes ( `` z '' and `` atoll '' ) in the colour - colour and colour - intensity plots .
these variations are primarily due to the main thermal component .
the temperature of the comptonizing medium ranges from @xmath4 kev ( `` soft '' state ) to @xmath5 kev ( `` hard '' state ) while corresponding electron scattering optical depth ranges from @xmath6 to @xmath7 ( e.g. * ? ? ?
the occurrence of khz qpo has been reported to be related to the spectral shape or more specifically to the position of the source in its colour - colour diagram @xcite .
the detection of the qpos in the long term light curves of sources have also been examined @xcite .
these studies seem to suggest that the qpos are preferentially detected in intermediate spectral states which lie between the extreme soft and hard states . on the other hand ,
statistically analysis undertaken by @xcite suggests that the fraction of segments having a khz qpo is near unity at low states and decreases to zero at high states with nearly constant amplitude . in any case , irrespective of the details mentioned above , the khz qpo is most probably associated with the primary thermal comptonization component which typically dominates the spectra .
the fractional root mean square ( r.m.s ) amplitude of the khz qpos have been measured to be an increasing function of energy @xcite with indications of a decrease at @xmath8 kev @xcite .
this again implies that the phenomenon should be associated with the high energy thermal comptonization component and not the soft black body one .
remarkable evidence was obtained when the time lag between different energy bins for the lower khz qpo was measured to be of the order of @xmath9 micro - seconds @xcite . for thermal comptonization ,
higher energy photons scatter more than the low energy ones and hence one expects time - lags between them . for a corona of size of few kms and optical depths of order unity ,
the expected time difference is of the order of @xmath10 micro - seconds .
it was realised that the variation of the r.m.s and the time - lag as a function of energy is a powerful diagnostic which can distinguish which parameter ( e.g. the plasma temperature , the soft photon input ) drives the oscillation as well as constrain the size and geometry of the system @xcite . for the lmxb 4u 1608 - 52
, the time lag was found to be `` soft '' i.e. the soft photons are delayed compared to the hard ones ( erratum of * ? ? ?
such soft lags were confirmed also for 4u 1636 - 53 @xcite .
this is in apparent contradiction to the thermal comptonization model , where one would expect hard lags .
however , @xcite showed that if a fraction of the comptonized photons impinge back in the soft photon source , then the system would show soft lags with an energy dependence similar to what has been observed .
earlier , time lags for khz qpos were reported for a single rxte observation of 4u 1608 - 52 @xcite and 4u 1636 - 53 @xcite .
recently @xcite and @xcite has analysed a number of observations of both sources and have confirmed that the time lag is soft and is of the order of tens of micro - secs for the lower khz qpo .
the time lag remains nearly constant till about @xmath11 hz and then decreases as the frequency increases .
@xcite have studied for the first time the time - lag for the upper khz qpo and find them to be hard lags and inconsistent with those seen for the lower khz qpo .
thus , these results ascertain that time - lags in khz qpos is a rich phenomenon which can provide important insight into the nature of these systems . apart from delays due to electron scatterings in a thermal plasma
, the time - lags could be due to other reasons .
longer time - lags ( @xmath12 msec ) observed in the low frequency continuum @xcite are attributed to propagation of stochastic variability in a disk @xcite or for a low frequency qpo ( e.g. * ? ? ?
* ) , time - lag may be due to non - linear dependence of spectral parameters with each other @xcite .
the micro - second time - lags observed for khz qpos are unlikely to be due to any of the above but could be due to reverberation of the x - rays from a reflector which is few kms away from the source @xcite .
however , it should be emphasised that since the primary spectral component is due thermal comptonization , time lag due to multiple electron scattering must occur and should be of the order of tens of micro - seconds . thus , while reverberation may contribute to the observed time - lags , the comptonization lag is an important effect , which can not be neglected for khz qpos . in this work ,
we study using the linearised time dependent kompaneets equation , the response of the thermal plasma to variations in the input photon flux and the plasma heating rate .
for the latter , we also consider the possibility that a fraction of photons impinge back into the input photon source .
the analysis predicts the shape of the r.m.s as a function of energy as well as the expected time - lags as a function of energy .
we apply the method to a test case and show that such analysis can provide constraints on the geometry and size of the system . in the next section
we develop the mathematical formulation which is followed by 3 where the generic results are presented . in
4 , as a test case , we apply the method to data from 4u 1608 - 52 to show the utility of such an analysis . in 5 we summarise and discuss the work .
thermal comptonization is a process in which low energy `` seed '' photons are comptonized by a hot thermal electron gas . while the geometry of the comptonizing system in an x - ray binary may be complex , here we assume a simple one , where a spherical seed photon source of radius @xmath13 is surrounded by a comptonizing medium of width @xmath14 and temperature @xmath15 .
the input source is considered to be a black body with temperature @xmath16 and hence the rate of input photons per unit volume inside the medium is given as @xmath17}\right ] \left ( \frac{2\pi}{h^3c^2 } \frac{e^2}{(\exp\left[{\frac{e}{kt_b}}\right]-1 ) } \right)\end{aligned}\ ] ] the thompson collision time scale is @xmath18 and the optical depth of the medium is @xmath19 , where @xmath20 is thompson cross - section and @xmath21 is the electron density .
the evolution of the photon density , @xmath22 , inside the comptonized medium in the non - relativistic limit ( i.e. kt@xmath23 @xmath24 m@xmath25 ) and for low photon energies ( e @xmath24 m@xmath25 ) is governed by the kompaneets equation @xcite where the induced scattering term is neglected : @xmath26 \nonumber \\ & + t_c\dot{n}_{s\gamma } - t_c\dot{n}_{esc}\end{aligned}\ ] ] it should be noted that eqn ( 2 ) describes the evolution of the photon density , @xmath22 , while typically the kompaneets equation is often written in terms of the photon occupation number @xmath27 . in this work
it is more intuitive to work directly with the photon density instead of the occupation number . here
@xmath28 is the rate of escape of the photon density which is taken to be @xmath29 @xmath30 . in steady state
, the equilibrium photon density @xmath31 can be computed by setting @xmath32 and @xmath33 represents the emergent spectrum .
the spectral shape is determined by the corresponding steady state values @xmath34 , @xmath35 and @xmath36 .
the electron temperature of the corona is set by the balance of compton cooling and some external heating .
in general the time evolution of the temperature is governed by @xmath37 where @xmath38 is the external heating rate per electron and the compton cooling rate per electron is given by @xmath39 in steady state , when the electron temperature is @xmath34 , @xmath40 .
we take into account the possibility that a fraction , @xmath41 , of the comptonized photons impinge back onto to the seed photon source , thereby increasing the temperature of the seed photon source which can be estimated by @xmath42 where @xmath43}$ ] is the volume of the comptonizing medium and @xmath44 is the input source temperature if @xmath45 i.e. @xmath46 is the energy generation rate inside the seed photon source .
if the variability amplitude of a source is small , then the time averaged spectrum will correspond to the steady state one with parameters , @xmath34 , @xmath47 and @xmath36 . for these values
, @xmath48 can be estimated to be @xmath49 and for the system there is a maximum allowed value of @xmath41 , @xmath50 for which @xmath51 .
we consider small amplitude oscillation of the medium electron temperature and the source photon temperature over their average values i.e. @xmath52 , and @xmath53 . here
, @xmath54 is the angular frequency of the oscillation and @xmath55 and @xmath56 are in general complex quantities . in the linear approximation
, these will lead to variation in the photon density as @xmath57 .
the oscillation amplitude of @xmath22 , @xmath58 is related to @xmath59 and @xmath60 by the linearised kompaneets equation , which is written in simplified form as ( e.g. * ? ? ?
* ; * ? ? ?
* ) @xmath61 here , @xmath62 is the time averaged input photon rate . solving eqn [ komlin ] one gets the complex variation @xmath63 as a function of energy , which can be used to compute the fractional r.m.s with energy , @xmath64 , as well as phase lag between two energies @xmath65 and @xmath66 which is the argument of @xmath67 $ ] .
we consider two possible drivers of the variability of the system ( i ) variation in the seed photon temperature @xmath16 and ( ii ) a variation in the external heating rate of the corona @xmath38 and analyse the induced temporal variation for both cases . a variation in the seed photon temperature
i.e. @xmath68 could be due to variation in the energy generation rate in the soft photon source .
this will lead to a variation in the photon number density @xmath58 as governed by eqn [ komlin ] .
the variation of the photon density would change the compton cooling rate , leading to a variation in medium temperature , @xmath59 which can be estimated by linearising eqn ( [ heat ] ) i.e. @xmath69 \label{heatlin}\end{aligned}\ ] ] where we have assumed that the external heating rate of the corona is a constant at the steady state value , @xmath70 .
the two complex differential equations ( [ komlin ] ) and ( [ heatlin ] ) have to be solved simultaneously to obtain @xmath59 and @xmath71 with the latter being a function of energy .
we use an iterative numerical scheme to solve them .
we start by having @xmath72 and obtain @xmath71 from eqn ( [ komlin ] ) .
this is then used to get an updated value of @xmath59 by eqn ( [ heatlin ] ) .
the procedure is iterated till @xmath59 converges .
for all computations done here , we find that the convergence is rapid and the converged value of @xmath59 is obtained within a few iterations . a variation of the coronal heating i.e. @xmath73
. would lead to a variation in the coronal temperature which is given by the linearised form of eqn ( [ heat ] ) i.e. @xmath74 \label{heatlin2}\end{aligned}\ ] ] the induced variation in the temperature would naturally lead to variations in the photon density .
there is additional complication because the photon density variation itself may lead to variations in the soft photon temperature , since a fraction of them are assumed to impinge back on to the soft photon source .
this effect can be estimated by linearizing eqn ( [ feedback ] ) , @xmath75 the three complex differential equations ( [ komlin ] ) , ( [ heatlin2 ] ) and ( [ feedbacklin ] ) have to be solved simultaneously to obtain @xmath59 , @xmath60 and @xmath71 .
the solutions can be readily obtained in an iterative manner which converges rapidly .
@xmath76{ffigure1l.pdf}&\hspace{-0.6 cm } \includegraphics[width=0.265\textwidth]{ffigure1r.pdf}\\ \end{array}$ ] @xmath76{ffigure2r.pdf}&\hspace{-0.6 cm } \includegraphics[width=0.265\textwidth]{ffigure2lcc.pdf}\\ \end{array}$ ] to illustrate and to understand the energy dependent variability of the comptonized photons , we consider two different steady state spectra .
the first is a `` hard '' spectrum ( model a ) characterised by relatively high temperature and low optical depth : @xmath77 kev , @xmath78 and a seed photon temperature of @xmath79 kev .
the second is a `` soft '' spectrum ( model b ) with low temperature and high optical depth , @xmath80 kev , @xmath81 and a seed photon temperature of @xmath82 kev . in practise , these time averaged values can be obtained by fitting the average spectrum with a steady state comptonization model .
here we use these fiduciary parameter values to bring out qualitative differences between any hard and soft average spectra . for both models
we choose the seed photon size @xmath83 km and the corona size to be @xmath84 km and the frequency of oscillation to be @xmath85 hz .
@xmath86{ffigure3_11.pdf}&\hspace{-0.6 cm } \includegraphics[width=0.265\textwidth]{ffigure3_12.pdf}\\\vspace{-2.8cm}\hspace{-0.35 cm } \includegraphics[width=0.265\textwidth]{ffigure3_21.pdf}&\hspace{-0.6cm}%######%{figure3_21c.eps}missing \includegraphics[width=0.265\textwidth]{ffigure3_22.pdf}\\\vspace{-2.8cm}\hspace{-0.35 cm } \includegraphics[width=0.265\textwidth]{ffigure3_31.pdf}&\hspace{-0.6 cm } \includegraphics[width=0.265\textwidth]{ffigure3_32.pdf}\\\vspace{0.80cm}\hspace{-0.35 cm } \end{array}$ ] @xmath86{ffigure4_11.pdf}&\hspace{-0.6 cm } \includegraphics[width=0.265\textwidth]{ffigure4_12c.pdf}\\\vspace{-1.6cm}\hspace{-0.35 cm } \includegraphics[width=0.265\textwidth]{ffigure4_21.pdf}&\hspace{-0.6 cm } \includegraphics[width=0.265\textwidth]{ffigure4_22c.pdf}\\ \end{array}$ ] before studying the dependence with energy , we consider the behaviour of the system for different oscillation frequencies .
the variation of the seed photon temperature leads to a variation of the corona temperature and the ratio @xmath87 has been plotted in the top panel of figure [ seedtemp_freq ] for the hard and soft spectra . at low frequencies
the coronal temperature responds to the seed photon variation with @xmath88 , however at higher frequencies , the coronal temperature is unable to react to the rapid variation and hence varies significantly less .
this is because at high frequencies the average photon density inside the corona fails to respond to rapid variation in the seed photons .
this is illustrated in the right panel of figure [ seedtemp_freq ] , where the energy integrated normalised photon density variation @xmath89 is plotted as a function of frequency .
@xmath90 is the normalised integrated photon density variation in the energy band @xmath91 - @xmath92 kev , i.e. @xmath93 . for both the `` soft '' and `` hard '' spectra , the photon density variation decreases sharply for frequencies larger than @xmath94 hz .
the photon density variation in the corona will lead to variations in the escaping photons and hence @xmath95 is proportional to the fractional r.m.s observed from a source .
thus , for the system under consideration , fluctuations with r.m.s @xmath96% is unlikely for frequencies @xmath97 hz , since that would imply large values of @xmath98 . for these curves ,
the size of the system is @xmath84 kms , and changing @xmath14 will lead to a proportional horizontal shift in frequency . for @xmath99 kms @xmath100 for a frequency of @xmath101 hz .
this implies that the observed khz qpo , restricts the size to be less than @xmath102 kms , which is similar to the constrain obtained by light travel time arguments of @xmath103 kms .
the energy dependence of the photon density variation @xmath104 is shown in figure [ seedtemp_energy ] for the `` hard '' and `` soft '' spectra .
@xmath58 is equal to the normalised r.m.s that can be obtained from observations .
the frequency of oscillation is @xmath85 hz . there is a pivot point where the photon variation goes to zero and it is at high energies @xmath105-@xmath106 kev . below the pivot point , the variation is nearly constant or decreasing with energy . also shown in figure [ seedtemp_energy ] is the expected lag between different energies .
the time lag is an increasing function of energy while implies that the hard photon lag the soft ones ( i.e. a hard lag ) .
@xmath107{lag03nn_52.pdf}&\hspace{-0.6 cm } \includegraphics[width=0.356\textwidth]{lag10nn_52.pdf}&\hspace{-0.6 cm } \includegraphics[width=0.356\textwidth]{lag20n_52.pdf}\\ \end{array}$ ] @xmath107{rms03nn_52.pdf}&\hspace{-0.6 cm } \includegraphics[width=0.356\textwidth]{rms10nn_52.pdf}&\hspace{-0.6 cm } \includegraphics[width=0.356\textwidth]{rms20n_52.pdf}\\ \end{array}$ ] variation of the coronal heating rate , @xmath108 will induce a variation of the coronal temperature , @xmath59 and their ratio as a function of frequency is plotted in figure [ cortemp_freq ] .
if there is a fraction of photons which impinge back into the soft photon source ( i.e. @xmath109 ) , then there will be a variation of the seed photon temperature as well and this is plotted in the middle panel of figure [ cortemp_freq ] . the bottom panel shows the energy integrated photon density variation @xmath110 as a function of frequency . for the figure , we have considered two extreme cases of @xmath45 and @xmath111 , where @xmath50 is the maximum allowed fraction . there is not much qualitative differences between the results of the `` soft '' and `` hard '' state spectra ( left and right columns of figure [ cortemp_freq ] ) which implies that the variability nature of such systems is not sensitive to the time averaged spectrum . however , there are important differences in the predicted variability of a system with @xmath45 and @xmath111 . for @xmath45
the energy integrated photon variability , @xmath112 is a constant and decreases for frequencies higher than @xmath94 hz , fairly similar to the results obtained for the case when the seed photon temperature is varied ( figure [ seedtemp_freq ] ) . for @xmath113 , @xmath112 starts decreasing for frequencies @xmath114 hz and is @xmath115 at @xmath1 khz . for any observed khz qpo with fractional r.m.s @xmath116% , would require an unphysically large variation in the heating rate @xmath117 .
for such a system to be viable its size should be much smaller than @xmath84 kms assumed here . note that this is a more stringent limit that that obtained from light travel time arguments of @xmath103 kms .
the variation of @xmath118 and time - lag as a function of photon energy are shown in figure [ cortemp_energy ] .
again , there are qualitative differences between the case when @xmath45 and @xmath119 . for @xmath45 ,
the pivot point is at low energies while no pivoting is seen for @xmath119 .
more strikingly , the time lag behaviour changes . while for @xmath45
, the hard photons lag the soft ones , the reverse occurs for @xmath119 .
thus , in this interpretation soft lags occur only when a significant fraction of the comptonized photons impinge back into the input photon source .
as discovered by @xcite , the rxte observation of the @xmath0 hz qpo of 4u 1608 - 52 on 3rd march 1996 , remains one of the best cases of soft lag detection in khz qpos .
recently , @xcite have estimated the time lag as a function of energy for more than 20 energy bins and have interpreted some of the qualitative features as due to reverberation lags .
thus , we choose this observation set as an example to illustrate how these time - lags can be interpreted in terms of comptonization time delays and the possible constraints that can be obtained . as shown in the previous section , soft time lags are expected only when the variation is in the coronal heating rate and with a significant fraction of photons impinging back into the input photon source .
a detailed analysis can be undertaken only when the different components of the time - averaged spectrum are well constrained .
in particular , the spectra of these sources require apart from a primary thermal component , a soft component modelled as a disk black body and a broad iron line , which implies the presence of a reflected component .
the @xmath120-@xmath121 kev low resolution spectrum obtained from the pca observations are not sufficient to constrain these complex parameters .
the absorption column density is undetermined and is often fixed to the galactic value which might be an underestimate . nevertheless , the primary component is the thermal comptonization and when we fit a single thermal component model to the time averaged spectrum of this observation , we find that the model fits the data with a systematic uncertainty of 10% .
the parameters obtained from such a fit , electron temperature @xmath122 kev , optical depth @xmath123 = 40 and soft photon temperature of @xmath124 kev are similar to those obtained from more sophisticated analysis ( e.g. * ? ? ?
we defer a detailed analysis taking into account different spectral parameters to a latter work .
apart from the spectral parameters , there are two other important parameters which determine the time - lag versus energy .
these are the size of the comptonizing region , @xmath14 and the fraction , @xmath41 of photons impinging back into the input source .
these two parameters can be constrained by comparison with the observed values .
the three panel in figure [ lag_1608 ] all show the observed time - lag versus energy for the observation which are compared with model predictions .
the curves in the left , middle and right panels correspond to three different sizes of the comptonizing regions , @xmath125 , @xmath126 and @xmath127 kms respectively .
the curves marked 1 , 2 , and 3 are for three different values of the fraction @xmath41 mentioned in the caption .
the best representation of the data occurs when @xmath128 km and @xmath129 ( solid line in the middle panel ) . for smaller values of @xmath125 kms ,
the magnitude of the time - lag is smaller than observed , while for larger values , @xmath130 kms it is larger .
thus we demonstrate that time - lag analysis can well constrain the size of the comptonizing region .
the model also predicts the r.m.s . versus energy for the qpo .
in such a linear model , the normalisation of the r.m.s .
scales with the magnitude of the perturbation @xmath59 , but the shape of the curve is determined by the spectral parameters , @xmath41 and the size @xmath14 of the system .
figure [ rms_1608 ] shows the variation of r.m.s versus energy for the same observation .
the data points have been taken from @xcite .
the lines in the three panels are the predicted r.m.s for the model parameters used in figure [ lag_1608 ] .
it is interesting to note that for photon energies @xmath131 kev , the predicted variation matches well with the observed values for @xmath132 km and @xmath129 ( solid line of second panel in figure [ rms_1608 ] ) .
the observed r.m.s . is lower for higher energies possibly due to the presence of an additional constant hard component .
the linearised time dependent kompaneets equation describing thermal comptonization has been solved self consistently taking into account energy balance of the plasma and soft photon source . when the perturbation is due to temperature variation in the soft photon source , the time - lag between photons of different energies is found to be hard .
the fractional r.m.s as a function of energy reveals a pivot point at high energies and is nearly constant at lower energies .
when the perturbation is in the heating rate of the plasma the pivot point is located at low energies .
there is a major qualitative change in the behaviour of the system if a significant fraction of the photons impinge back into the soft photon source . in this case
, the time - lag is found to be soft and the r.m.s increases monotonically with energy .
such behaviour has been reported for the khz oscillations in lmxbs . as an example , we compare the model predictions with the time - lag and r.m.s observed for the 850 hz oscillation of 4u 1608 - 52 and show that the size of the comptonizing region can be constrained to @xmath1 km .
the analysis shows that modelling the time - lag and r.m.s variation as a function of energy for khz qpos is a powerful technique which can in principle constrain the size and geometry of the source .
a more detailed analysis will require a careful modelling of the time - dependent spectra .
modelling the time - lag as well as the r.m.s at different qpo frequency and taking into account the corresponding spectral changes , may provide information regarding how the size and geometry of the system changes . moreover , while the time - lags for the upper khz qpo have not been constrained , they are known to be significantly different than those of the lower khz qpo and hence should provide additional constraints .
we defer such a detailed analysis to explain the observations reported by @xcite to a latter work .
one may also need to take into account additional effects such as reverberation lags .
the iron line component and the soft disk emission may oscillate in response to the plasma variation after a time - lag . however , it is important to note that even for relatively small plasma size of @xmath133 km , the time - lags due to comptonization is of the order of @xmath10 microseconds and hence they are expected to dominate .
the more subtle and complicated general relativistic effects may also be needed to be incorporated in order to correctly model such systems . while the qualitative results such as the presence of soft lags do not depend on the time averaged spectral parameters , the constraints obtained on the size and geometry
do depend on their precise values .
thus , it is important to obtain broad band spectral data during the time when a khz qpo has been observed .
this will be possible by the upcoming multi - wavelength satellite astrosat @xcite where the large array x - ray proportional counter will detect the qpo while it and the other x - ray instruments will provide broad band spectra .
berger , m. , van der klis , m. , van paradijs , j. , lewin , w.h.g . , lamb , f. , vaughan , b. , kuulkers , e. , augusteijn , t. , zhang , w. , marshall , f.e . ,
swank , j.h .
, lapidus , i. , lochner , j.c . ,
strohmayer , t.e . , 1996 , apj , 469 , l13 | we study the energy dependent photon variability from a thermal comptonizing plasma that is oscillating at khz frequencies . in particular , we solve the linearised time dependent kompaneets equation and consider the oscillatory perturbation to be either in the soft photon source or in the heating rate of the plasma . for each case , we self consistently consider the energy balance of the plasma and the soft photon source .
the model incorporates the possibility of a fraction of the comptonized photons impinging back into the soft photon source .
we find that when the oscillation is due to the soft photon source , the variation of the fractional root mean sqaure ( r.m.s ) is nearly constant with energy and the time - lags are hard .
however , for the case when the oscillation is due to variation in the heating rate of the corona , and when a significant fraction of the photons impinge back into the soft photon source , the r.m.s increases with energy and the time lags are soft . as an example , we compare the results with the @xmath0 hz oscillation observed on march 3 , 1996 for 4u 1608 - 52 and show that both the observed soft time - lags as well as the r.m.s versus energy can be well described by such a model where the size of the comptonizing plasma is @xmath1 km .
thus , modelling of the time lags as due to comptonization delays , can provide tight constraints on the size and geometry of the system .
detailed analysis would require well constrained spectral parameters . [ firstpage ] stars : neutron x - rays : binaries x - rays : individual : 4u 160852 radiation mechanisms : thermal |
understanding the spin structure of the nucleon is one of the most important questions in strong interaction physics today .
there has been renewed interest in recent years , largely as a result of a series of high precision polarized deep inelastic scattering experiments .
so far , the bulk of the information comes from measurements of the spin - dependent structure function @xmath0 . from such measurements ,
spin - dependent parton distributions can be extracted and compared with theoretical models , or used to make predictions for future experiments .
several theoretical issues have received particular attention : the predictions of various sum rules and the measurement of @xmath1 , the magnitude of the polarized gluon distribution , and the behaviour of the polarized parton distributions at small @xmath2 . in a previous study @xcite
we performed a global analysis of all available deep inelastic polarized structure function data in the context of the qcd improved parton model at leading order .
we supplemented the experimental information with theoretical assumptions about the flavour content and the form of the distributions at large and small @xmath2 .
in particular , we presented three different sets of parton distributions characterized by qualitiatively different polarized gluon distributions , and suggested further experiments which could discriminate between them . since the work of ref .
@xcite there have been several important advances .
more precise polarized structure function data has become available . both the the smc @xcite and slac
e143 collaborations @xcite have measured the structure function @xmath3 of the deuteron , supplementing their previous proton target measurements , and the slac
e143 collaboration has also measured the structure function @xmath4 on proton and deuteron targets @xcite .
the new @xmath3 data are particularly important for determining the polarized @xmath5 and @xmath6quark distributions separately .
a second advance has been the theoretical calculation @xcite of the next to leading order contributions to the polarized splitting functions @xmath7 which determine the @xmath8 evolution of the distributions .
this not only improves the precision of the phenomenology , but also allows for the first time a consistent factorization / renormalization scheme dependence to be imposed on the analysis . in this study
we repeat the analysis of @xcite using the new polarized structure function data , but now at both leading and next to leading order in perturbation theory . our aim , as before , is to derive a consistent set of polarized parton distributions . at the same time we can check that the @xmath8 dependence seen in the data is consistent with the predictions of perturbative qcd .
we discuss in detail the constraining power of the various data sets , the role of the polarized gluon distribution , and the prospects for future measurements .
a similar analysis of polarized structure functions has been reported recently in ref .
@xcite , using the ` dynamical parton model ' approach in which the distributions at small @xmath2 are generated dynamically from valence like distributions at a small starting @xmath9 scale .
a reasonable description of the data is obtained .
our approach , in contrast , is to constrain the small @xmath2 distributions from the data , without prejudice as to the origin of the observed behaviour .
this is in the same spirit as the unpolarized analysis of refs .
@xcite .
we begin by reviewing the basic theoretical input to the analysis .
the fundamental quantity of interest is the structure function @xmath3 which , in analogy with the unpolarized structure function @xmath10 , can ( in the qcd improved parton model at leading order ) be expressed in terms of the probability distributions for finding quarks with spin parallel or antiparallel to the longitudinally polarized parent nucleon : @xmath11 \\ g_1(x , q^2 ) & = & \half \sum_q\ ; e_q^2\ ; [ \delta q(x , q^2 ) + \delta\bar q(x , q^2 ) ] , \label{naiveg1}\end{aligned}\ ] ] where @xmath12 the first moment of @xmath0 measures the expectation value of the axial
vector current between two identical nucleon states : @xmath13 where the conserved nonsinglet axial vector currents are known from the hyperon decay constants ( the ellis jaffe sum rule @xcite ) : @xmath14 in contrast , the singlet axial vector current @xmath15 is non conserved due to the axial anomaly .
it can not be related to any other experimental quantity , although several model dependent estimates of it can be found in the literature : if the total nucleon spin is carried only by quarks one finds @xmath16 @xcite , and if the strange quark sea in the nucleon is unpolarized then @xmath17 @xcite .
both these estimates are clearly ruled out by experimental measurements @xcite , which find @xmath18 in the range @xmath19 . in the qcd improved parton model @xmath3
is expressed in terms of parton distributions for the polarization of quarks and gluons , @xmath20\ ; \nonumber \\ & & \times \left\{\delta(1-y ) + { \alpha_s(q^2 ) \over 2\pi } \delta c_q(y ) + \ldots \right\ } \nonumber \\ & & + \ ; \frac{\langle e_q^2 \rangle}{2 } \int_x^1 \ ; { dy\over y}\ ; \delta g(x / y , q^2 ) \ ; \left\ { n_f\ ; { \alpha_s(q^2 ) \over
2\pi}\;\delta c_g(y ) + \ldots \right\}. \label{qcdg1}\end{aligned}\ ] ] the variation of these polarized parton distributions with @xmath8 is determined by the glap evolution equation @xcite : @xmath21 for a consistent theoretical description of @xmath0 , the coefficient functions in ( [ qcdg1 ] ) must be truncated at the same order in perturbation theory as the splitting functions in ( [ evo ] ) .
although the @xmath22 corrections to the coefficient functions @xmath23 @xcite and @xmath24 @xcite have been known for a long time , a next to leading order study was not possible until very recently , as the @xmath22 corrections to the @xmath25 were not known .
these have now been published for the first time in ref .
@xcite . using the next to leading order evolution equations @xcite , the _
unpolarized _ parton distributions can be determined to a high level of accuracy ( see for example refs .
@xcite ) using a wide variety of high precision experimental data from different processes . as the only data available on the spin structure of the nucleon are the structure function measurements of refs .
@xcite , we are still far from obtaining a similar level of precision on the polarized distributions .
the appropriate choice of scheme for @xmath3 in the qcd improved parton model has been a matter of discussion over the last few years ( see for example @xcite for a recent review ) . the main issue is the treatment of the gluonic contributions to the ellis
jaffe sum rule .
while the nonsinglet contributions to this sum rule can be related to conserved currents and are unambiguously defined , there is an arbitrariness in the decomposition of the singlet contribution @xcite , depending on the absorption procedure used for the axial anomaly . in the @xmath26 scheme
, this anomaly is absorbed into the next to leading order pure singlet quark to quark splitting function @xcite .
another possible procedure is to allow for an anomalous gluonic contribution to the ellis
jaffe sum rule , which is absent in the @xmath26-scheme .
the appropriate scheme transformation @xcite then yields conservation of the quark contribution to @xmath15 . in the present study
we will work entirely in the @xmath26 scheme .
any other choice of scheme would tie the resulting distributions to @xmath3 , as higher order corrections to the coefficient functions of other processes are so far only known in the @xmath26 scheme . and
the first moment of the polarized gluon distribution . ] the choice of the @xmath26 scheme also allows us to use our distributions in conjunction with a broad range of unpolarized distributions , the majority of which are defined in this scheme . for reference ,
the @xmath26 scheme coefficient functions in ( [ qcdg1 ] ) are @xmath27 in the following section we will present the results of our leading and next to leading order parton model analyses of polarized structure function data .
section 3 discusses the prospects of various future experiments on polarized nucleons and section 4 contains a brief summary and our conclusions .
we adopt a similar approach to the global analysis of unpolarized parton distributions in the nucleon @xcite by parametrizing the polarized distributions at the starting scale in the form : @xmath28 where we take @xmath29 gev@xmath30 .
the normalization factors are @xmath31 which ensures that the first moments of the distributions , @xmath32 , are given by @xmath33 . various experimental measurements of unpolarized lepton(@xmath34)nucleon and unpolarized drell yan cross sections yield a reasonably precise flavour decomposition of the light quark @xmath35 sea .
such a decomposition is not yet possible for the polarization of the light quark sea , as measurements of the structure function @xmath3 are only sensitive to the charge weighted sum of all quark flavours , not to the individual distributions .
we therefore assume a su(3)symmetric antiquark polarization @xmath36 .
this _ ad hoc _
assumption is only justified at the present level of experimental knowledge , and is furthermore immediately broken by next to leading order evolution @xcite .
the first moments of the polarized quark distributions can be determined from the measured values of the ellis
jaffe sum rule . imposing su(3)symmetry at @xmath9 ,
this sum rule reads @xmath37 with @xmath38 in this approach , the first moments of the valence quark polarizations are obtained from the nonsinglet axial
vector current matrix elements @xcite , while the first moment of the sea quark distribution is inferred from the measured value of @xmath39 . for the leading order ( lo ) distributions
, we correct the normalization of @xmath40 and @xmath41 by the @xmath42 coefficient function in ( [ eq : gamma1 ] ) @xcite .
the first moments obtained by this procedure are listed in table [ tab : fixed ] .
note that the ellis jaffe sum rule is a ( @xmath8independent ) constant at leading order in perturbation theory , as the leading - order coefficient functions are only expanded up to @xmath43 and scaling violations in @xmath44 arise only from the splitting functions at next to leading order .
the polarized gluon distribution enters @xmath0 at next to leading order .
it is only very weakly constrained so far , as no experimental data are available on gluon initiated processes such as direct@xmath45 or heavy meson production .
the polarized gluon distribution @xmath46 is therefore not well determined by a fit to the @xmath3 data alone , and so additional theoretical constraints have to be applied .
it follows from the structure of the polarized splitting functions that the small-@xmath2 behaviour of the gluon and sea quark distributions are closely correlated , which justifies the assumption @xmath47 in ( [ pform ] ) . in the region @xmath48 ,
structure functions and their evolution are dominated by valence quark contributions , and the impact of the gluon is completely negligible . in ref .
@xcite we explored various possibilities for the form of @xmath46 at large @xmath2 : hard and soft distributions with the spin aligned with that of the parent hadron , and a distribution with the spin anti aligned .
all three choices give equally good descriptions of the structure function data , but would be relatively easy to discriminate if data on polarized gluon initiated processes were available .
we adopt the same procedure here , i.e. we consider three equally possible scenarios for the behaviour of @xmath49 , which can be parametrized as follows : @xmath50 due to the introduction of the additional parameter @xmath51 in the starting parametrizations ( [ pform ] ) , these distributions look slightly different from the ones presented in @xcite .
the normalization @xmath52 of the gluon distribution can only be determined consistently from the experimental data in a next to leading order analysis , where it still has a large error . at leading order
, we can estimate @xmath52 by attributing _ all _ the violation of the ellis
jaffe sum rule to a large gluon polarization and vanishing sea quark polarization . in this way
we obtain @xmath53 , only slightly different from the value 1.971 obtained in ref .
note , however , that at leading order the apportioning of the singlet contribution to @xmath39 between gluons and sea quarks is completely arbitrary @xcite .
in fact we shall see below that a consistent nlo treatment gives a value of @xmath52 somewhat less than our estimated leading order value , and similar to the range of values found in ref .
@xcite .
if parton distributions are interpreted in the probabilistic picture of the naive parton model , the magnitude of the polarized distributions can not exceed the unpolarized distributions , in order to guarantee positive probabilities for the individual polarization states , i.e. @xmath54 this is in fact only a rigid constraint at leading order , since parton distibutions at higher orders are only scheme
dependent constants of renormalization and not strict probability densities .
the fundamental constraint at arbitrary orders in perturbation theory is the positivity of physical cross sections for all possible helicity configurations , which does not necessarily imply the positivity of the distributions .
positivity of the polarized distributions is achieved by constraining the parameters of the starting distributions at @xmath9 such that @xmath55 perturbative evolution preserves the positivity of the individual helicity distributions , hence ( [ pos ] ) is fulfilled at any @xmath8 . in our leading
order analysis , we use the unpolarized distributions from @xcite for reference . at @xmath56
these are @xmath57 ( 1-x)^{6.356 } \nonumber\\ xs ( x , q_0 ^ 2 ) & = & 0.0034 ( -\ln x)^{-1.15 } ( 1 - 2.392x^{1/2}+7.094x ) ( 1-x)^{6.166 } \nonumber \\ & & \exp \left(\sqrt{-6.719\ln x}\right ) \nonumber \\ xg ( x , q_0 ^ 2 ) & = & \big [ x^{0.731}(5.110 - 1.204x-1.911x^2 ) ( -\ln x)^{-0.4718 } + \nonumber\\ & & 0.0527 \exp \left(\sqrt{-4.584\ln x}\right ) \big ] ( 1-x)^{5.566}\end{aligned}\ ] ] the reference unpolarized distributions at next - to - leading order are the a@xmath58 set of @xcite , which are parametrized at @xmath59 as @xmath60 note that the choice of unpolarized distributions is not crucial for the present analysis .
all the widely available leading and next to leading order distributions provide very good fits to the unpolarized structure function data , and the small differences between them are much smaller than the precision with which the polarized distributions are currently determined .
the flavour decomposition of the unpolarized sea quark distributions is also unimportant for our present analysis and will therefore be disregarded .
the starting sea quark distribution of ref .
@xcite contains a very small charm quark contribution which can safely be ignored in the present analysis . to be consistent with the evolution of the unpolarized distributions ,
we take @xmath61 these correspond to @xmath62 ( lo ) and @xmath63 ( nlo ) .
the parameters most affected by the positivity constraints are the large-@xmath2 exponents @xmath64 . for the valence quarks ,
we fix @xmath65 and @xmath66 , motivated by counting rule estimates @xcite .
the parameters @xmath67 and @xmath68 are constrained to be at least as large as their unpolarized counterparts in the fit , but this constraint has only minimal impact .
we find that only the @xmath69 distribution tends to saturate positivity , requiring the combination @xmath70 to be limited in the fit .
the data currently available on @xmath3 are not able to test the various theoretical model predictions for the small-@xmath2 behaviour of the polarized parton distributions @xcite , which are only expected to apply at much lower values of @xmath2 @xcite .
the parameters @xmath71 are therefore only effective exponents valid over some finite interval in @xmath2 .
it therefore makes no sense to postulate positivity for @xmath72 by constraining the @xmath71 . ' '' '' & lo & nlo + ' '' '' @xmath40 & 0.823 & 0.918 + ' '' '' @xmath41 & @xmath730.303 & @xmath730.339 + ' '' '' @xmath52 & 1.9 & + ' '' '' @xmath74 & @xmath730.0495 & @xmath730.060 + ' '' ''
@xmath75 & 3.73 & 3.96 + ' '' '' @xmath76 & 4.73 & 4.96 + the contribution of charmed quarks to @xmath0 is negligible at present experimental energies @xcite and will not be considered in this analysis .
we therefore adopt the evolution procedure of ref .
@xcite and fix the number of flavours in the splitting functions at @xmath77 , while the number of flavours in @xmath1 increases at each mass threshold , @xmath78 and @xmath79 is determined by requiring @xmath1 to be continuous across each threshold . rather than measuring @xmath0 directly from absolute cross section differences , it is the relative asymmetry @xmath80 which is determined experimentally .
the structure function @xmath0 is then inferred using a particular parametrization of @xmath81 .
some experimental groups assume @xmath8scaling of @xmath82 in their extraction of @xmath0 . in order to have a consistent set of data
, we have instead used the measured values of @xmath82 quoted by the experiments and re - evaluated @xmath0 from eq .
( [ a1def ] ) , constructing @xmath81 from the parametrizations of @xmath83 @xcite and @xmath84 @xcite which were used in the most recent measurements . applying the constraints outlined above
, we have used all available world data on @xmath85 @xcite to fit the polarized quark and gluon distributions with the parametric forms of ( [ pform ] ) imposed at @xmath86 .
about 35% of these data were taken at @xmath87 . to have sensible constraints on the distributions , in particular for @xmath88
, we include these datapoints in the global fit .
the distributions in the region @xmath89 are obtained by inverting the evolution matrix , which is straightforward in @xmath90-moment space .
a problem with using low @xmath8 data points in the fit is the possible contamination by higher
twist contributions .
we have tried to estimate to magnitude of such contributions to @xmath3 using the parametrization of @xmath91 from ref .
@xcite and assuming @xmath92 .
the higher twist contributions estimated in this way are found to be small for all data points apart from the two lowest @xmath2 bins of the smc experiment .
the global fit is performed using glap evolution algorithms in @xmath90-moment space @xcite .
the distributions and structure functions are then restored by a numerical inversion into @xmath2 space .
the results of the global fit using the leading and next to leading order @xmath93 expressions for the splitting functions and the @xmath3 coefficient functions are listed in table [ tab : res ] .
the resulting distributions at @xmath9 are shown in figs . 1 ( lo ) and 2 ( nlo ) . ' '' '' & a ( lo ) & b ( lo ) & c ( lo ) & a ( nlo ) & b(nlo ) & c ( nlo ) + ' '' '' @xmath94 & 0.578 & 0.585 & 0.582 & 0.512 & 0.504 & 0.471 + ' '' ''
@xmath95 & 9.38 & 9.31 & 9.50 & 11.65 & 11.98 & 13.14 + ' '' '' @xmath96 & @xmath734.26 & @xmath734.28 & @xmath734.28 & @xmath734.60 & @xmath734.61 & @xmath734.90 + ' '' '' @xmath97 & 0.666 & 0.662 & 0.660 & 0.780 & 0.777 & 0.809 + ' '' ''
@xmath98 & 10.46 & 10.91 & 11.04 & 7.81 & 8.18 & 6.73 + ' '' '' @xmath99 & @xmath735.10 & @xmath735.09 & @xmath735.06 & @xmath733.48 & @xmath733.61 & @xmath731.99 + ' '' '' @xmath52 & & & & 1.71 & 1.63 & 1.02 + ' '' '' @xmath100 & 0.520 & 0.524 & 0.456 & 0.724 & 0.670 & 0.425 + ' '' ''
@xmath67 & 9.45 & 6.87 & 8.72 & 5.71 & 5.34 & 11.05 + ' '' '' @xmath101 & 15.06 & 15.96 & 11.82 & 14.40 & 18.06 & 16.40 + ' '' '' @xmath102 & 2.30 & 2.42 & 2.11 & 4.63 & 5.30 & @xmath732.67 + ' '' '' @xmath103 & @xmath732.00 & @xmath732.00 & @xmath731.95 & @xmath734.96 & @xmath735.25 & @xmath733.08 + ' '' '' @xmath104 & 98.3 & 97.7 & 100.0 & 89.7 & 91.0 & 93.4 + the resulting parameters are not independent of each other . in particular , @xmath94 , @xmath97 and @xmath105 are strongly correlated .
the @xmath106 of the valence distributions are anticorrelated with the corresponding @xmath107 , reflecting the fact that @xmath106 is only an effective exponent for a finite range in @xmath2 .
the @xmath107 and @xmath108 are also anticorrelated .
the @xmath104 distribution is very flat around the local minima found by the global fits , especially with respect to the gluon and sea quark parameters .
the three gluon scenarios give fits of almost identical quality , reflecting the small impact of the gluon distribution on @xmath0 at large and medium @xmath2 .
the @xmath104 obtained in the nlo fits are systematically lower due to the additional degree of freedom given by the normalization of the polarized gluon distribution .
all fits give very good descriptions for the polarized structure functions @xmath109 .
this is illustrated in fig .
3 , which shows the nlo description of the various @xmath3 measurements using gluon a. the curves correspond to @xmath110 gev@xmath30 , reflecting the spread in @xmath8 values of the different data sets .
there is a systematic decrease in the @xmath8 values of the data points from large @xmath2 to small @xmath2 .
the contributions of @xmath111 and @xmath112 to the neutron structure function @xmath113 are almost equal in magnitude but opposite in sign .
the neutron structure function is therefore much more sensitive to the sea quark polarization than @xmath114 and @xmath115 .
it displays a clear double peak structure , as the sea quarks are dominant in a different @xmath2-region than the valence quarks
. a precision measurement of @xmath113 @xcite will therefore be able to provide vital new information on the shape of the sea quark polarization . from a consideration of the size of the errors on the various fitted parameters ,
it is apparent that the world data on @xmath0 really only constrain the polarized valence quark distributions and , to a lesser extent , the overall magnitude of the sea quark polarization .
the flavour decomposition of the polarized sea is still completely unknown
. only dedicated experiments , such as the production of drell yan lepton pairs or the flavour tagging of final state hadrons in polarized deep inelastic scattering , will be able to provide further information .
most important of all , the polarized gluon distribution is almost completely undetermined , as its impact on the polarized structure function is less than the present experimental accuracy .
the variation between our three gluon sets certainly underestimates the true uncertainty in the distributions .
there are several measurements which could be feasible at future experiments and which yield additional information on the polarized gluon distribution . in this section
we present some representative predictions for two of these : the @xmath8 dependence of @xmath0 , and an asymmetry in @xmath116photoproduction .
we also briefly discuss the prospects of a measurement of polarized structure functions at the hera collider .
the derivative of @xmath117 with respect to @xmath8 has been used to measure the unpolarized gluon distribution in fixed target experiments @xcite ( @xmath118 ) and at hera @xcite ( @xmath119 ) .
the method is particularly powerful at small @xmath2 , where the gluon distribution dominates the @xmath8 evolution , @xmath120 . in the same way
, we can use our three sets of distributions a , b and c to explore the sensitivity of the polarized structure function evolution to @xmath46 .
figure 4 shows the predictions for the asymmetry @xmath82 as a function of @xmath8 in the kinematic range representative of current fixed target experiments .
the asymmetry is obtained from eq .
( [ a1def ] ) with @xmath3 calculated using our polarized distributions and @xmath10 calculated using the nlo unpolarized mrs(a@xmath58 ) distributions of ref .
the latter are extrapolated to lower values of @xmath8 , which reproduces the full backwards evolution to within a few per cent . at large @xmath2 , there is no sensitivity to @xmath46 the evolution is completely dominated by the quark contribution . at small @xmath2 , on the other hand , we see some dependence on the gluon . in this region @xmath121 , and
so the derivative @xmath122 is more negative for the sets that have a larger gluon polarization above the @xmath2 value considered . in particular
, we see that at @xmath123 the proton asymmetry is almost @xmath8 independent for set c , but decreases with increasing @xmath8 for sets a and b. unfortunately , the sensitivity of the present experiments is much worse than the differences between the various sets . to illustrate this , we have included two data points at @xmath124 from the recent smc and e143 measurements . for the deuteron ,
the quark contribution to the structure function is smaller , and so the dependence on @xmath125 at small @xmath2 is somewhat enhanced . considering the large errors on the present data ,
it seems doubtful that a measurement of the polarized gluon distribution from the @xmath8 variation of @xmath82 is feasible for values of @xmath8 where perturbative expressions can be safely applied .
future high statistics experiments at higher beam energies @xcite will clearly improve this measurement .
it still has to be kept in mind that a determination of @xmath125 from the @xmath8 dependence of @xmath82 can never reach the quality of the corresponding unpolarized measurement , as the gluonic contribution is not as dominant in the evolution of polarized parton distributions as it is in the unpolarized evolution @xcite .
inelastic @xmath116 production from a nucleon target directly probes the gluon distribution via the photon
gluon fusion subprocess @xmath126 , which can be described in the colour singlet model of ref .
@xcite . depending on the virtuality of the photon , one can distinguish two different classes of events : photoproduction @xmath127 and leptoproduction @xmath128 .
the emc @xcite and nmc experiments @xcite have obtained measurements of the _ unpolarized _ leading order gluon distribution in the range @xmath129 from this process .
the results agree well with gluon distributions extracted from other processes , for example large@xmath130 direct photon production .
the corresponding cross section for _ polarized _
@xmath116 leptoproduction has been calculated in ref .
@xcite . taking the photoproduction limit @xmath131 ,
one obtains @xmath132 ^ 2 } \nonumber \\ & & \times\ [ a(z)+h\lambda c(z ) ] \;\cdot \ ; { \cal f}(p_t^2 , z ) \ ; , \end{aligned}\ ] ] with @xmath133 the polarized ( unpolarized ) cross sections are obtained by taking the difference ( sum ) over the helicity states .
the polarized cross section is then proportional to @xmath134 and depends only on @xmath135 , while the unpolarized cross section is proportional to @xmath136 and depends only on @xmath137 .
taking the cross sections integrated over one of the variables , one can define two physics asymmetries : @xmath138 with a photon beam of energy @xmath139 gev and cuts on @xmath140 and @xmath141 ( @xmath142 gev@xmath30 , @xmath143 , as proposed in @xcite ) , one gets significantly different asymmetries for the three gluon distributions a , b and c. predictions for the asymmetry @xmath144 , calculated using the three leading order distributions in combination with the unpolarized leading order gluon distributions of ref .
@xcite , are shown in fig .
note that small @xmath141 corresponds to large @xmath145 , and the ordering of the asymmetry in this region simply reflects the ordering of the different @xmath146 at large @xmath145 .
due to the relatively large asymmetry of up to 15% , such a measurement could provide vital information on the leading order polarized gluon distribution for @xmath147 and @xmath148 . in the foreseeable future
, the hera collider may be able to accelerate polarized protons @xcite .
this would offer the unique opportunity of measuring the polarized structure function @xmath114 far beyond the @xmath8 range of present fixed target experiments . in order to judge the quality of the new information gained from such a measurement ,
an order of magnitude estimate of the statistical errors is crucial @xcite .
the structure function @xmath114 is extracted from a measurement of the asymmetry @xmath149 where @xmath150 ( @xmath151 ) denote the polarizations of the electron ( proton ) beam and @xmath152 is the kinematical depolarization of the photon .
the statistical error on @xmath153 is therefore @xcite : @xmath154 & = & \frac{1}{2\lambda_e \lambda_p } \frac{1+(1-y)^2}{1-(1-y)^2 } \left[2xf_1^p(x , q^2 ) + \frac{2(1-y)}{1+(1-y)^2 } f_l^p(x , q^2)\right ] \nonumber\\ & \times & \left ( { \cal l}_{int } \int [ \d^2 \sigma^{\rm ( unpol.)}/ ( \d x \d q^2 ) ] \d x \d q^2 \right)^{-1/2 } \sqrt{1-a_{\parallel}(x , q^2)},\end{aligned}\ ] ] where the unpolarized differential cross section is integrated over the bin used in the measurement . to study the accuracy of a measurement of @xmath155 in the collider mode of hera , we have evaluated the above expression using the unpolarized parton distribution set mrs(a@xmath58 ) from ref .
@xcite with the next to leading order polarized parton distribution set a described in section 2 .
we apply the following cuts to the hera phase space @xcite : @xmath156 , @xmath157 , @xmath158 and @xmath159 , and consider two scenarios for the beam energies : @xmath160 the expected errors for an integrated luminosity of @xmath161 at @xmath162 in these two scenarios are shown in fig .
we take two bins per decade in @xmath2 and only one bin in @xmath8 .
these results are consistent with the leading order estimates of @xcite , bearing in mind the different cuts applied .
in particular , we estimate smaller errors in the small@xmath2 region , as we apply no cuts on the hadroinc final state .
the average @xmath8 values probed at hera range from @xmath163 ( @xmath164 ) to @xmath165 ( @xmath166 ) for @xmath167 and from @xmath168 ( @xmath169 ) to @xmath170 ( @xmath166 ) for @xmath171 . for reference , we also show in fig . 6 a parametrization of @xmath114 obtained from the nlo set a. to illustrate the impact of a measurement at hera , we include the three lowest datapoints reported by the smc experiment @xcite , corresponding to @xmath8 values around @xmath172 .
it is apparent that a measurement at lower beam energies will yield data of higher statistical quality . in contrast
, the higher beam energies yield a measurement at smaller @xmath2 .
a measurement of @xmath114 at the hera collider will evidently not provide a large number of precision data on the @xmath2 and @xmath8 dependences of this structure function .
hence , it will not provide sufficient information for an indirect determination of the polarized gluon density .
the important physics result of such a measurement is the determination of the small@xmath2 behaviour of @xmath114 , for which various , significantly different predictions exist @xcite .
it is important to stress that a measurement of @xmath114 at small @xmath2 will reduce the experimental uncertainty on the ellis
jaffe sum rule .
the impact of the small @xmath2 region can be easily seen in fig .
6 , as the ellis
jaffe sum rule is proportional to the area enclosed by @xmath173 and the @xmath2 axis .
a final remarkable point on the measurability of @xmath114 at hera is the impact on the statistical error of the minimum cut on @xmath174 . as the photon depolarizes for small @xmath174 , a small cut on @xmath174 ( such as @xmath175 ) diminishes the statistical accuracy , even though more data are included in the bin .
we find that a minimal cut on @xmath174 between 0.1 and 0.2 yields the optimal accuracy .
we have performed leading and next to leading order qcd fits to the world data on the @xmath3 polarized structure function measured with proton , neutron and deuteron targets .
the quality of the fits is excellent , and from them we obtain sets of polarized parton distributions which can be used for further phenomenological analyses .
the experimental precision is highest for the proton and deuteron data , and together these constrain the shapes of the valence @xmath5 and @xmath6 distributions .
the sea quark and gluon distributions are still largely undetermined .
there is a weak constraint on the overall size of the former , but almost no information at all on the flavour decomposition of the sea .
following our earlier work , we have presented three qualitatively different gluon distributions , characterized by different behaviours at large @xmath2 .
it seems that further progress in the determination of polarized parton distributions will require measurements other than @xmath3 at fixed
target experiments .
inelastic @xmath116 photoproduction in lepton
hadron scattering has already proved capable of measuring the unpolarized gluon distribution , and a similar measurement with a polarized beam and target would provide the first real information on @xmath46 .
a measurement of the @xmath8 dependence of @xmath82 can help to estimate the overall amount of gluon polarization at higher and medium @xmath2 .
althogh present experiments can not provide such a measurement due to low statistics or too low values of @xmath8 , first information can be expected from the next generation of fixed target experiments .
we have also studied the accuracy of a measurement of @xmath0 at the hera collider with polarized beams .
such a measurement will help to discriminate between various predictions for the behaviour of @xmath0 at small @xmath2 , and will reduce the uncertainty on the ellis jaffe sum rule from the extrapolation to small @xmath2 .
financial support from the uk pparc ( wjs ) , and from the gottlieb daimler- und karl benz - stiftung and the studienstiftung des deutschen volkes ( tg ) is gratefully acknowledged .
this work was supported in part by the eu programme `` human capital and mobility '' , network `` physics at high energy colliders '' , contract chrx - ct93 - 0357 ( dg 12 coma ) .
we would like to thankl alan martin , werner vogelsang and dick roberts for useful discussions .
10 t. gehrmann and w.j .
stirling , z. phys .
* c65 * ( 1995 ) 461 .
smc collaboration : d. adams et al .
lett . * b357 * ( 1995 ) 248 .
slac - e143 collaboration : k. abe et al .
* 74 * ( 1995 ) 346 ; * 75 * ( 1995 ) 25 .
slac - e143 collaboration : k. abe et al .
, slac - pub-95 - 6982 ( 1995 ) .
r. mertig and w.l .
van neerven , leiden preprint inlo - pub-6/95 ( revised ) ( 1995 ) .
+ w. vogelsang , rutherford laboratory preprint ral - tr-95 - 071 ( 1995 ) .
m. glck , e. reya and w. vogelsang , phys . lett .
* b359 * ( 1995 ) 201 .
+ m. glck , e. reya , m. stratmann and w. vogelsang , dortmund preprint do - th 95 - 13 ( 1995 ) .
martin , r.g .
roberts and w.j .
stirling , phys .
b354 * ( 1995 ) 155 .
cteq collaboration : h.l .
lai et al .
* d51 * ( 1995 ) 4763 .
m. gourdin , nucl .
b38 * ( 1972 ) 418 . j. ellis and r.l .
jaffe , phys . rev .
* d9 * ( 1974 ) 1444 ; erratum
* d10 * ( 1974 ) 1669 .
slac - yale collaboration : m.j .
alguard et al .
* 37 * ( 1976 ) 1261 ; g. baum et al . ,
* 45 * ( 1980 ) 2000 ; * 51 * ( 1983 ) 1135 .
+ emc collaboration : j. ashman et al .
, nucl . phys . *
b328 * ( 1989 ) 1 . + slac - e142 collaboration : d.l .
anthony et al .
* 71 * ( 1993 ) 959 .
+ smc collaboration : b. adeva et al .
, phys . lett . *
b302 * ( 1993 ) 553 ; d. adams et al .
, phys . lett . *
b329 * ( 1994 ) 399 .
ahmed and g.g .
ross , nucl .
* b111 * ( 1976 ) 441 .
+ k. sasaki , progr .
* 54 * ( 1975 ) 1816 .
+ g. altarelli and g. parisi , nucl
b126 * ( 1977 ) 298 .
j. kodaira , s. matsuda , t. muta , k. sasaki and t. uematsu , phys .
* d20 * ( 1979 ) 627 . j. kodaira , nucl . phys . *
b165 * ( 1980 ) 129 .
e.g. floratos , d.a .
ross and c.t .
sachrajda , nucl . phys . * b129 * ( 1977 ) 66 ; errata * b139 * ( 1978 ) 545 , * b152 * ( 1979 ) 493 .
+ a. gonzales arroyo , c. lopez and f.j .
yndurain , nucl .
* b153 * ( 1979 ) 161
. + g. curci , w. furmanski and r. petronzio , nucl . phys .
* b175 * ( 1980 ) 27 , + w. furmanski and r. petronzio , phys . lett .
* b97 * ( 1980 ) 437 .
+ e.g. floratos , c. kounnas and r. lacaze , nucl . phys .
* b192 * ( 1981 ) 417 .
m. glck , e. reya and a. vogt , z. phys . * c67 * ( 1995 ) 433 .
m. anselmino , a. efremov and e. leader , phys . rep .
* 261 * ( 1995 ) 1 .
lam and bing - an li , phys . rev .
* d25 * ( 1982 ) 683 .
zijlstra and w.l .
van neerven , nucl
b417 * ( 1994 ) 61 ; erratum * b426 * ( 1994 ) 245 .
ball , s. forte and g. ridolfi , preprint cern th/95266 ( 1995 ) .
ross and c.t .
sachrajda , nucl .
b149 * ( 1979 ) 497 .
+ m. stratmann , w. vogelsang and a. weber , dortmund preprint do - th 95 - 15 ( 1995 ) . -266
close and r.g .
roberts , phys .
b316 * ( 1993 ) 165 .
brodsky , m. burkardt and i. schmidt , nucl .
* b441 * ( 1995 ) 197 .
ahmed and g.g .
ross , phys .
* 56b * ( 1975 ) 385 .
close and r.g .
roberts , phys .
b336 * ( 1994 ) 257 .
+ s.d . bass and p.v .
landshoff , phys . lett .
* b336 * ( 1994 ) 537 . + r.d .
ball , s. forte and g. ridolfi , nucl .
b444 * ( 1995 ) 287 ; erratum * b449 * ( 1995 ) 680 .
+ j. bartels , b.i .
ermolaev and m.g .
ryskin , desy preprint desy-95 - 124 ( 1995 ) .
+ j. blmlein and a. vogt , desy preprint desy-95 - 175 ( 1995 ) .
+ j. kwieczinski , durham preprint dtp/95/98 ( 1995 ) , to appear in acta physica polonica .
t. gehrmann and w.j .
stirling , durham preprint dtp/95/62 ( 1995 ) , to appear in phys .
lett . * b*. m. glck , e. reya and w. vogelsang , nucl .
* b351 * ( 1991 ) 579 .
nmc collaboration : p. amaudruz et al .
, phys . lett .
* b295 * ( 1992 ) 159 .
whitlow et al .
, phys . lett . *
b282 * ( 1992 ) 159 . m. virchaux and a. milsztajn , phys . lett .
* b274 * ( 1992 ) 221 .
w. furmanski and r. petronzio , z. phys .
* c11 * ( 1982 ) 293 .
hermes collaboration : technical design report , desy ( 1993 ) .
nmc collaboration : m. arneodo et al .
, phys . lett . *
b309 * ( 1993 ) 222 .
h1 collaboration : s. aid et al .
, phys . lett . *
b354 * ( 1995 ) 494 .
+ zeus collaboration : m. derrick et al .
, phys . lett . *
b345 * ( 1995 ) 576 .
slac - e154/e155 collaborations : proposals ( 1994 ) .
berger and d. jones , phys . rev . *
d23 * ( 1981 ) 1521 .
+ r. baier and r. rckl , nucl . phys . *
b218 * ( 1983 ) 289 .
emc collaboration : j. ashman et al . , z. phys . *
c56 * ( 1992 ) 21 .
nmc collaboration : d. allasia et al .
, phys . lett . *
b258 * ( 1991 ) 493 .
guillet , z. phys .
* c39 * ( 1988 ) 75 .
v. breton , _ measurement of @xmath176 by @xmath116 photoproduction at slac _ , slac proposal , january 1994 .
proceedings of the `` workshop on the prospects of spin physics at hera , august 28 - 31 , 1995 , desy - zeuthen , zeuthen , germany '' , eds . j. blmlein and w.d .
nowak , desy report , to appear .
j. blmlein , desy preprint desy-95 - 164 .
m. klein , proceedings of the workshop `` physics at hera '' , hamburg , october 2930 , 1991 , eds .
w. buchmller and g. ingelman , p. 71
* leading order polarized parton distributions as described in the text at @xmath177 compared to the unpolarized distributions of . * next to leading order polarized parton distributions as described in the text at @xmath177 compared to the unpolarized distributions of . * structure function measurements @xmath153 of the proton , deuteron and neutron compared to the next to leading order predictions obtained using gluon a. * @xmath8 dependence of @xmath178 in next to leading order using the gluons a , b and c. to illustrate the sensitivity of current experiments , we show the @xmath179 datapoints from the recent smc and e143 measurements . * predictions for the asymmetry @xmath137 ( eq . ( [ eq : az ] ) ) in polarized @xmath116 photoproduction at @xmath180 , using the different leading order polarized gluon distributions .
* expected errors for a measurement of @xmath114 at the hera collider with @xmath181 and @xmath182 . | the distribution of the spin of the nucleon among its constituents can be parametrized in the form of polarized parton distribution functions for quarks and gluons . using all available data on the polarized structure function @xmath0 , we determine these distributions both at leading and next to leading order in perturbation theory .
we suggest three different , equally possible scenarios for the polarized gluon distribution , which is found to be only loosely constrained by current experimental data . we examine various possibilities of measuring polarized parton distributions at future experiments .
20.5 cm dtp/95/82 + december 1995 + 1.cm * polarized parton distributions in the nucleon * 1.cm t. gehrmann and w.j . stirling .4 cm _ departments of physics and mathematical sciences , university of durham + durham dh1 3le , england _ + |
spherically symmetric potentials have become a well - established paradigm of colloidal science in past decades .
@xcite this is because , at a sufficiently coarse - grained level , colloidal surface composition can be regarded as uniform with a good degree of confidence , so that relevant interactions depend only on relative distances among the particles .
recent advances in chemical particle synthesis @xcite have however challenged this view by emphasizing the fundamental role of surface colloidal heterogeneities and their detailed chemical compositions .
this is particularly true for an important subclass of colloidal systems , namely proteins , where the presence of anisotropic interactions can not be neglected , even at the minimal level .
@xcite directional interactions introduce novel properties in such systems .
these properties depend both on the number of contacts ( _ i.e. _ , the valency ) and the amplitude of these interactions ( _ i.e. _ , flexibility of the bonds ) , a notable example of this class being hydrogen - bond interactions , ubiquitous in biological , chemical , and physical processes .
@xcite as a reasonable compromise between the high complexity of interactions governing the above systems and the necessary simplicity required for a minimal model , patchy - sphere models stand out for their remarkable success in this rapidly evolving field .
@xcite see ref . for a recent review on the subject . within this class of models
, interactions are spread over a limited part of the surface , either concentrated over a number of pointlike spots @xcite or distributed over one or more extended regions .
@xcite while the former have the considerable advantage of a simple theoretical scheme @xcite which allows a first semi - quantitative description , the latter can easily account for both the effect of the number of contacts and their amplitude , unlike `` spotty '' interactions which are always limited by the one - bond - per - site constraint .
in this paper we consider a particular model due to kern and frenkel @xcite of this patchy - spheres class wherein short - range attractive interactions of the square - well ( sw ) form are distributed over circular patches on otherwise hard spheres ( hs ) .
interactions between particles ( spheres ) are then attractive in the sw - sw interfacial geometry or purely hard - sphere repulsive under the hs - sw or hs - hs interfacial geometries , and can sustain more than one bond in fact , as many as the geometry allows even in the case of a single patch assigned to each sphere . a number of real systems ranging from surfactants to globular proteins can be described with simplified interactions of these particular forms , with well - defined solvophilic and solvophobic regions , and despite their simplicity patchy hard spheres have already shown a remarkable richness of theoretical predictions .
@xcite notwithstanding the discontinous nature of the angular interactions , highly simplified integral equation approaches are possible , @xcite but only very recently has a complete well - defined scheme , within the framework of the reference hypernetted - chain ( rhnc ) integral equation , been proposed and solved for patchy spheres .
@xcite this integral equation belongs to a class of approximate closures which have been extensively exploited in the field of molecular associating fluids .
@xcite its main advantage over other available approximations ( other than its less - accurate parent hnc closure ) lies in the fact that it relies on a single approximation , for the bridge function appearing in the exact relation between pair potential and pair distribution function @xmath4 , @xcite to directly yield structural and thermodynamic properties that include the helmholtz free energy and the chemical potential with no further approximations .
@xcite in addition , it can be made to display enhanced consistency among different thermodynamic routes .
@xcite this is an important point when analyzing fluid - fluid phase diagrams such as we propose to do here .
we thus build upon our previous work with the one - patch potential @xcite to study the two - patch case and its relationship with its one - patch counterpart .
in addition to rhnc integral equation results , we provide dedicated monte carlo simulations which can assess the performance of rhnc .
we find that rhnc provides a robust representation of both structural and thermophysical properties of the two - patch kern - frenkel model for a wide range of coverage @xmath0 ( the ratio between attractive and total hard - sphere surface ) , extending from an isotropic sw to a bare hs potential .
the competition arising between phase separation and polymerisation is discussed in terms of the angular dependence of the pair correlation function and the structure factor .
finally , a comparison between the one - patch and two - patch phase diagrams shows a strong impact on the different morphology and stable structures obtained in the two cases .
we also report numerical simulation results of the model in the region where the rhnc integral equations do not numerically converge , to explore the low temperature , small @xmath0 limit .
we find that for @xmath5 it becomes impossible to investigate the low - temperature disordered phases , since the system quickly transforms into an ordered structure , which itself depends on the coverage value .
indeed , on decreasing @xmath0 one progressively enters the region where the maximum number of contacts per patch evolves from four to two and eventually reaches the one - bond - per - patch condition .
when three or four bonds per patch are possible , the observed ordered structure is a crystal of interconnected planes , while when only two contacts are possible , particles order themselves into a set of disconnected planes . the patchy interaction model examined here can be regarded as a prototype of a special colloidal architecture where there exist competitive interactions on the colloidal surface that drive , by free energy minimization , the different colloidal particles through a spontaneous self - assembly process into complex superstructures whose final target can be experimentally probed and properly tuned
. @xcite the possibility , discussed in the present study , of identifying the position of the gas - liquid coexisting lines and its relative interplay with different structures , opens up fascinating scenarios in material science , on the possibility of novel material design exploiting a bottom - up process not requiring human intervention .
as a paradigmatic model for highly anisotropic interactions , we take the kern - frenkel @xcite two - patch model where two attractive patches are symmetrically arranged as polar caps on a hard sphere of diameter @xmath6 .
each patch can be reckoned as the intersection of a spherical shell with a cone of semi - amplitude @xmath7 and vertex at the center of the sphere .
consider spheres @xmath8 and @xmath9 and let @xmath10 be the direction joining the two sphere centers , pointing from sphere @xmath8 to sphere @xmath9 ( see fig .
[ fig : fig1 ] ) .
the orientation of sphere @xmath11 is defined by a unit vector @xmath12 passing outward through the center of one of its patches , to be arbitrarily designated as the `` top '' @xmath13 patch .
the patch on the opposite , `` bottom '' @xmath14 pole is then identified with the outward normal @xmath15 .
two spheres attract via a square - well potential of range @xmath16 and depth @xmath17 if any combination of the two patches on each sphere are within a solid angle defined by @xmath7 and otherwise repel each other as hard spheres .
the pair potential then reads @xcite @xmath18 where @xmath19 and @xmath20 where @xmath21 or @xmath22 indicates which patch , top or bottom , is involved on each sphere .
the unit vectors @xmath23 are defined by the spherical angles @xmath24 in an arbitrarily oriented coordinate frame and @xmath25 is identified by the spherical angle @xmath26 in the same frame .
reduced units , temperature @xmath27 and density @xmath28 , will be used throughout .
this model was introduced by kern and frenkel , @xcite patterned after a similar model studied by chapman _
et al . _ , @xcite as a minimal model where both the distributions and the sizes of attractive surface regions on particles can be tuned . in this sense
, the model constitutes a useful paradigm lying between spherically symmetric models which do not capture the specificity of surface groups , not even at the simplest possible level , and models with highly localized interactions having the single - bond , single - site limitation .
@xcite several previous studies have already examined potentials of the kern - frenkel form using numerical simulations , @xcite corresponding - state arguments , @xcite highly simplified integral equation theories , @xcite and pertubation theories .
@xcite more recently , @xcite the single patch kern - frenkel potential was studied using a more sophisticated integral equation approach based on the rhnc approximation coupled with rather precise and extensive monte carlo simulations . in the present paper
, we extend this last study to the two - patch kern - frenkel potential and provide new methodologies specific for the angular distribution analysis .
we define the coverage @xmath0 as the fraction of the total sphere surface covered by attractive patches .
thus @xmath29 corresponds to a fully symmetric square - well potential while @xmath30 corresponds to a hard - sphere interaction and the model smoothly interpolates between these two extremes in the intermediate cases @xmath31 .
the two - patch potential is expected to present qualitative as well as quantitative differences with respect to its one - patch counterpart .
one interesting question , for instance , concerns the subtle interplay between distribution and size of the attractive patches on the fluid - fluid phase separation diagram .
it is now well established @xcite that as coverage decreases the fluid - fluid coexistence line progressively diminishes in width and height .
indeed , this feature can be exploited to suppress phase separation altogether to enhance the possibility of studying glassy behavior @xcite and can not be accounted for with a simple temperature and density rescaling , @xcite although corresponding - state type of arguments can be proposed .
@xcite on the other hand , the above mechanism can significantly depend on how the same reduced attractive region is distributed on the surface of particles . in ref .
, for instance , it was suggested that lines of decreasing critical temperature as a function of decreasing coverage , for the one - patch and two - patch kern - frenkel models with very short - range interactions , could cross each other at a specific coverage : for low coverages , critical temperatures for the one - patch model lie above the two - patch counterpart whereas the opposite is true for larger coverages .
this would have far - reaching consequences on the phase diagram , as phase separation would occur at higher or lower temperatures for fixed coverage , depending on the specific allotment of the coverage .
another interesting issue regards micellization phenomena , present in the one - patch version of the model , @xcite which is expected to be replaced by polymerization ( or chaining ) in the two - patch version . @xcite
the ornstein - zernike ( oz ) equation @xcite defines the direct correlation function @xmath32 in terms of the pair correlation function @xmath33 ; it is convenient for computation to write it using the indirect correlation function @xmath34 instead of @xmath35 .
we have then @xmath36 c\left(32\right).\end{aligned}\ ] ] a second , or `` closure , '' equation coupling @xmath37 and @xmath32 is needed . the general form for this
is @xcite @xmath38 - 1-\gamma\left(12\right),\end{aligned}\ ] ] where @xmath39 and a third pair function , the so - called `` bridge '' function @xmath40 , has also been introduced . while known in a formal sense as a power series in density , @xcite @xmath40 can not in fact be evaluated exactly and at this point an approximation is unavoidable .
the rhnc approximation replaces the unknown @xmath40 with a known version @xmath41 from some `` reference '' system . in practice ,
only the hard - sphere model is today well - enough known to play the role of reference system . here
we will use the verlet - weis - henderson - grundke parametrization @xcite for @xmath42 , where @xmath43 is the reference hard - sphere diameter .
some computational details of the rhnc integral equation approach can be found in ref .
( see expecially appendix a ) , so only the most relevant equations will be repeated here .
solution of the ornstein - zernike integral equation for molecular fluids @xcite seemingly requires expansions in spherical harmonics of the angular dependence of all pair functions , a need that would be very problematic in the case of the discontinuous angular dependence in the present @xmath44 .
in fact , the integral equation algorithm allows @xmath44 to remain unexpanded .
@xcite there is a potential problem however in evaluating the gauss - legendre quadratures used in the numerical solution , in that of the angles @xmath45 used for an @xmath46th - order quadrature , none is likely to coincide with the angle @xmath7 defining the semi - amplitude of a patch .
thus the algorithm will not `` know '' the correct patch size .
this problem is ameliorated in the following _ ad hoc _ fashion . from the interaction @xmath44 of eq .
( [ ie : eq3 ] ) , the total coverage @xmath0 can be computed in terms of @xmath7 as @xmath47,\end{aligned}\ ] ] where @xmath48 is the heaviside step function , equal to @xmath8 if @xmath49 and @xmath50 if @xmath51 .
the integrals can be readily evaluated to give @xcite @xmath52 this quantity can also be numerically evaluated by gauss - legendre quadrature using the @xmath46 roots @xmath53 of the legendre polynomial @xmath54 and the computed result compared with the exact value ( [ rhnc : eq5 ] ) .
we may then vary @xmath46 so as to find that number @xmath46 ( typically kept between 30 and 40 ) that minimizes the known error in computing @xmath0 .
all gaussian quadratures for that @xmath0 value are then evaluated with the same number @xmath46 of points , thus ensuring that minimal error arises from the selected angular grid . in an axial * r * frame @xcite with @xmath55 , the internal energy per particle in units @xmath56 is obtained from @xmath57 where @xmath58 denotes an average over spherical angle @xmath59 and where we have written out @xmath60 .
similarly , the pressure @xmath61 is computed from the compressibility factor @xmath62 \right \rangle_{\omega_1 \omega_2 } \right \},\end{aligned}\ ] ] where the cavity function @xmath63 has been introduced .
the angular integrations in these expressions are evaluated with gauss - legendre and gauss - chebyshev quadratures .
finally , the dimensionless free energy per particle @xmath64 and chemical potential @xmath65 can also be directly computed from the pair functions produced by the rhnc equation ; the overall calculation is optimized by choosing the reference hard sphere diameter @xmath43 so as to minimize the free energy functional .
@xcite we solve the rhnc equations numerically on @xmath66 and @xmath67 grids of @xmath68 points , with intervals @xmath69 and @xmath70 , using a standard picard iteration method .
@xcite the square - well width is set at @xmath71 as a reasonable value dictated by the availability of isotropic square well results .
@xcite further details of these and other computations can be found in ref . . for an assessment of the performance of the rhnc integral equation ,
we also perform nvt , grand canonical , and gibbs ensemble monte carlo ( mc ) simulations @xcite following the path set in the one - patch case . @xcite standard nvt mc simulations of a system of 1000 particles are used to compute structural information ( pair correlation functions and structure factors ) for comparison with integral equations results , whereas grand canonical and gibbs ensemble mc ( gemc ) are used to locate critical parameters and coexisting phases .
the exact locations of the critical points ( points connected by the thick dashed green line in fig . [ fig : fig2 ] ) have been obtained from the mc data assuming the ising universality class and properly matching the density fluctuations with the known fluctuations of the magnetization close to the ising critical point .
@xcite for gemc , we use a system of 1200 particles , which partition themselves into two boxes whose total volume is @xmath72 , corresponding to an average density of @xmath73 . at the lowest temperature
considered , this corresponds to roughly @xmath74 particles in the liquid box and 150 particles in the gas box ( of side @xmath75 ) . on average ,
the code attempts one volume change every five particle - swap moves and 500 displacement moves .
each displacement move is composed of a simultaneous random translation of the particle center ( uniformly distributed between @xmath76 ) and a rotation ( with an angle uniformly distributed between @xmath77 radians ) around a random axis .
we have studied systems of size @xmath78 up to @xmath79 to estimate the size dependence of the critical point , with an average of one insertion / deletion step every 500 displacement steps in the case of grand canonical monte carlo ( gcmc ) .
we have also performed a set of gcmc simulations for different choices of @xmath80 and @xmath81 to evaluate @xmath82 .
see ref . and additional references therein for details .
locating coexistence lines is not an easy task within integral equation theory , given the fact that virtually all integral equations are unable to access the critical region with reliable precision due to significant thermodynamic inconsistencies among various possible routes to thermodynamics , a consequence of the approximation buried in the closure eq .
( [ rhnc : eq2 ] ) .
the rhnc closure is no exception to this rule , but has the strong advantage of relying on a single approximation expressed by the choice of the reference bridge function @xmath41 , at odds with other available closures which require additional approximations in constructing various thermodynamic quantities such as the chemical potential .
here we follow the protocol outlined in ref . for the isotropic square - well potential and ref . for the one - patch kern - frenkel potential , where both the well - known pseudo - solutions shortcoming @xcite and the numerical drawbacks @xcite can be conveniently accounted for .
figure [ fig : fig2 ] depicts the location of the fluid - fluid coexistence line for the two - patch case upon varying the coverage @xmath0 .
the limiting case @xmath29 corresponds to the square - well potential . both mc ( points ) and rhnc ( thick solid lines ) results are shown . as previously noted , rhnc is not able to approach the critical point close enough to provide a direct estimate of its location .
however , since it provides a quite good description of the low - temperature part of the coexistence line , we have tried to use these data to approximately locate the critical point .
visual inspection of the rhnc coexistence points reveals , in the cases where it is possible to go closer to the critical region , an unphysical change of curvature of the coexistence line moving from low to high temperature .
for this reason , for each coverage we selected only data clearly consistent with a rectilinear diameter law .
then we fitted those data with the following function ( corresponding to the first correction to the scaling @xcite ) : @xmath83 where the values of the exponents @xmath84 and @xmath85 are appropriate for the 3d ising model universality class .
@xcite once @xmath86 and the amplitudes @xmath87 , @xmath22 have been determined , the critical density can be obtained from the rectilinear diameter best fit .
the numerical results of such a procedure are compared with mc estimates of the critical points in table [ tab : tab1 ] .
it is evident that even though in general the validity of eq .
( [ correction ] ) is deemed to be limited to a smaller neighborhood of the critical point , @xcite in the present case it provides an acceptable procedure for a quick first estimate of the critical point location .
as the coverage decreases , the coexistence line shrinks and moves to lower temperature and density , as expected from an overall - decreasing attractive interaction .
this trend can be tracked rather precisely by mc simulations down to remarkably low coverages ( @xmath88 ) and rhnc correctly reproduces this evolution down to @xmath89 coverage . below this value ,
more powerful algorithms are required to achieve good numerical convergence .
a few remarks are here in order .
the coexistence curves shown in fig .
[ fig : fig2 ] are consistent with previous analogous results reported in ref . but extend the range of temperatures and , more importantly , the range of coverages ( @xmath0 values ) .
this allows a quantitative measure of the significant deviation from the simple mean - field - like results which can be obtained from the simple scaling ( not shown ) @xmath90 as suggested by the second - virial coefficient @xmath91 for this model , @xcite @xmath92 @xmath93 being the hard - sphere result .
this is also consistent with the breakdown of the above simple scaling at the level of the third virial coefficient , derived in ref . for the companion patchy sticky - hard - sphere model . as we shall see later on ,
the dependence of the critical temperature and density on _ both _ the coverage _ and _ the number of patches is one of the main results of the present work . as a final point
, we note that all curves in fig .
[ fig : fig2 ] collapse into a single master curve upon scaling @xmath94 , in agreement with ref . .
below @xmath88 , it becomes impossible to properly estimate the location of the critical point or the density of the coexisting gas and liquid phases .
indeed , the gas - liquid separation becomes pre - empted by crystallization into a structure that depends on the value of @xmath0 .
hence , the liquid phase , as an equilibrium phase , ceases to exist for small @xmath0 .
this is strongly reminiscent of the disappearence of the liquid phase using spherical potentials when the range of the interaction becomes smaller than about 10% of the particle diameter , @xcite thus providing the angular analog of the same phenomenon .
interestingly enough , the crystal structure which is spontaneously observed during the simulation depends on the value of @xmath0 , since the @xmath0 value controls the maximum number of bonds per patch . in the range of @xmath0 values such
that each patch can be involved in four bonds , the observed ordered structure is made by planes exposing the sw parts to their surfaces ( see fig .
[ fig : fig3](d ) ) . particles in the plane are located on a square lattice and adjacent planes are shifted in each direction by a half lattice constant , resulting in a reduced energy of -4 per particle ( _ i.e. _ , eight bonded neighbors ) . on decreasing @xmath0 below @xmath95 , the region where only three bonds per particle are possible ( @xmath96^{-1 } < \sin\theta_0 < [ \sqrt{2}(1+\delta/\sigma)]^{-1}$ ] ) is entered and the crystal structure is made by interconnected planes of particles arranged in a triangular lattice ( see fig .
[ fig : fig3](c ) ) . for @xmath97^{-1 }
< \sin\theta_0 < [ \sqrt{3}(1+\delta/\sigma)]^{-1}$ ] only two bonds per patch are possible and the system organizes into independent planes ( see fig . [
fig : fig3](b ) ) , this time turning a hs surface to their neighboring planes . particles in the plane are now arranged on a triangular lattice and each patch is able to bind only to two different neighbors located in the same plane , resulting in a reduced energy per particle of @xmath98 . when @xmath99 becomes smaller than the value @xmath100^{-1}$ ] ( corresponding to @xmath101 )
, the one - bond - per - patch condition is reached and the system can form only isolated chains ( see fig .
[ fig : fig3](a ) ) . in this limit
, the system is expected to behave as the two single - bond - per - patch model .
we turn our attention next to structural information , where the advantages of a reliable integral equation approach become evident .
one has to keep in mind that gibbs ensemble and gcmc simulations are particularly painstaking , due to the combined effect of the required low temperatures and the aggregation properties of the fluid ( as detailed below ) , so that many of the state points examined here require several weeks of computer time . on the other hand , the rhnc integral equation , while rather demanding from an algorithmic point of view ( see _ e.g. _ appendix a of ref . )
is a rapidly convergent scheme yielding solutions on the order of minutes , depending on the temperatures considered .
a more profound advantage stems from the fact that , within the approximation defined by the rhnc closure , all possible pair structural information is in fact exactly available , unlike mc calculations where , though available in principle , their statistics would be so limited as to make such calculations impractical .
thus , only the pair correlation function @xmath4 averaged over angle @xmath25 is computed . by symmetry , the resulting pair function in this context depends only on @xmath102 and @xmath103 and will be denoted here as @xmath104 ; see appendix b in ref . for details .
consider then the unaveraged pair correlation function @xmath60 from rhnc in an axial frame with @xmath55 .
two noteworthy configurations occur when ( a ) all four patches lie along the same line ( we denote this as the parallel @xmath105 configuration , with @xmath106 , the actual labeling of each patch being unimportant ) and when ( b ) patches on sphere @xmath9 lie on an axis perpendicular to those of sphere @xmath8 ( in this case @xmath107 , which we denote as the crossed ( x ) configuration ) .
figure [ fig : fig4 ] reports the results for the case @xmath71 , @xmath108 , and @xmath109 , which has been selected so that the fluid is above the coexistence line for all considered coverages , with configurations @xmath110 and x in the top and bottom panels respectively .
values of coverages range from a full square - well potential ( @xmath29 ) to a hard - sphere potential ( @xmath30 ) .
as coverage decreases , the contact value @xmath111 of the @xmath110 configuration has the unusual behavior of first a slight increase from @xmath29 to @xmath112 , followed by a more marked increase starting at @xmath113 up to the very small coverage @xmath114 limit which eventually backtracks to roughly the same value as at @xmath113 in the hard - sphere limit . at the opposite side of the well , @xmath115 ,
an even more erratic behavior is observed , with an increase in the range @xmath116 , then a decrease for @xmath117 , a new increase down to @xmath114 , and a final sudden decrease to the hard sphere value @xmath30 .
a somewhat similar feature occurs in the x configuration where within the entire well region @xmath118 one observes a sudden decrease from @xmath29 to @xmath119 and a more gradual increase until reaching the highest value for the hs case .
it is worth noting that in the x configuration there is no discontinuous jump at @xmath120 for any value @xmath121 .
the reason for this has already been addressed in ref .
for the one - patch case .
outside the first shell , there is a very weak dependence on the coverage , with a slight shift in the location of the second peak from a value of @xmath122 at the sw @xmath29 to a value of @xmath123 for lower @xmath0 .
we compare rhnc integral equation results with mc simulations in fig .
[ fig : fig5 ] .
as noted above , only the averaged pair function @xmath104 can be compared and this is done in the figure for different values of the coverage @xmath0 at the same state point ( @xmath124 and for the same @xmath110 and x configurations considered earlier . the good overall performance of rhnc in representing mc results is apparent as both contact values at the well edges and the jump discontinuities are very well reproduced .
it is instructive to contrast these results with those of fig .
[ fig : fig4 ] , as many of the abrupt changes appearing in the actual pair correlation function are smoothed out by the orientational average carried out here . for instance ,
the characteristic jump at @xmath120 of the @xmath110 configuration ( top panel ) progressively decreases as coverage is reduced and disappears in the hard - sphere limit .
conversely , the jump is present also in the x configuration ( bottom panel ) unlike the corresponding case of the full @xmath4 .
in addition , the strong increase of the @xmath110 configuration for low coverages is not present in this figure ; as remarked earlier , this level of detail in @xmath4 is one of the main advantages of an integral equation approach .
an additional useful quantity to consider , in view of its direct experimental access through scattering experiments , @xcite is the structure factor , which will be denoted @xmath125 within our theoretical framework .
@xcite this is also strongly related , via hankel transforms , to the radial distribution function @xmath126 , which is @xmath4 averaged over all orientations @xmath127 and @xmath128 of the patches and of the relative angular position @xmath26 .
note that @xmath126 is also the simplest rotational invariant ( see ref . and
appendix [ app : appb ] ) . the structure factor and the radial distribution function
are reported in fig .
[ fig : fig6 ] for two representative values of coverage , @xmath129 and @xmath130 , corresponding to almost fully attractive and almost fully repulsive limits .
these values have been selected at the same state point previously considered ( @xmath109 , @xmath108 ) as having a very different behavior within the first shell @xmath131 .
this high - density result is also contrasted with a low - density state point @xmath132 at the same temperature , a value which , in the temperature - density plane , lies symmetrically with respect to the coexistence curves in the single fluid phase for all coverages ( see fig .
[ fig : fig2 ] ) .
a few features are worth noting . for density
@xmath108 there is a significant coverage dependence , where the contact value @xmath133 for @xmath130 coverage is larger than that for the corresponding @xmath129 case and , conversely , the jump present at the other extreme @xmath134 is much smaller in the former than in the latter case .
a similar feature also occurs for the low - density state point @xmath132 .
this results from an angular average of the results given in fig .
[ fig : fig4 ] .
likewise , there is a marked difference in the behavior of the structure factor @xmath125 for the high - density case @xmath108 , both in the height of the first peak ( related to the @xmath133 value ) and of the secondary peaks ( related to the behaviors of @xmath126 in the @xmath135 region and of the @xmath136 discontinuity ) .
similarly , in the low - density branch @xmath132 , the large @xmath137 value for the @xmath129 coverage case is signaling the approach to a spinodal instability which is clearly not present in the corresponding @xmath130 coverage
. one natural interpretation of the above results is the progressive rearrangement of the distribution within the first shell upon varying both the coverage and the density . to support this view
, we consider the angular distribution within the first shell in the next subsection . the nonmonotonic dependence of @xmath4 in terms of the distance @xmath138 for decreasing coverage @xmath0 , as illustrated in fig .
[ fig : fig4 ] , is rather intriguing and requires an explanation . a similar ,
albeit different , feature occurs even in the one - patch case , as shown in ref . .
we have tackled this in two ways , illustrated in the following .
all previous representations of @xmath4 have been depicted in the molecular axial frame , where @xmath139 , so that patches on sphere @xmath8 are parallel to the vector @xmath140 joining sphere @xmath8 with sphere @xmath9 .
this is clearly preventing an understanding of the angular distribution of the patches around a given sphere @xmath8 , that is , as a function of @xmath25 .
this is however a needless restriction , as one can start from the expression for @xmath4 in a general ( laboratory ) frame and study the dependence on the angle @xmath26 for fixed patch directions @xmath141 and @xmath142 . in figs .
[ fig : fig4 ] , [ fig : fig5 ] and [ fig : fig6 ] , we notice that the main dependence on coverage @xmath0 stems from the region within the well , @xmath131 ; it sufficies therefore to investigate the _ average _ @xmath26 dependence by integrating over the radial variable @xmath66 within this region .
we further note that there is azimuthal symmetry with respect to the @xmath143 variable , so that we can focus on the @xmath144 dependence .
the details of the analysis are reported in appendix [ app : appa ] , where it is shown that the relevant quantity is @xmath145 , which is a function of the angle @xmath144 ( polar dependence of @xmath25 ) and of the polar angle @xmath146 of the patches on particle @xmath9 , given that the patches on particle @xmath8 lie along the @xmath147 axis .
we report comparative calculations for both low ( @xmath132 ) and high ( @xmath108 ) density at identical temperature @xmath109 at two representative coverages , @xmath129 , representing a case with almost all attraction , and @xmath130 , as representative of an almost hard - sphere case .
these are the same conditions considered in fig .
[ fig : fig6 ] ; the results are reported in fig .
[ fig : fig7 ] .
let us consider first the high - density , @xmath108 , situation as depicted in the two left panels for @xmath129 ( a ) and @xmath130 ( c ) . here
, the @xmath129 case yields a very well - defined pattern with a periodically modulated distribution of the patches in symmetrical fashion as indicated by the trimodal distribution as a function of the relative positional angle @xmath144 , so that @xmath148 are almost equally represented .
( note that @xmath149 are necessarily equivalent due to the up - down symmetry of the two - patch distribution . )
the two interstitial minima are a consequence of the reduced valency the corresponding fully symmetrical result under this condition would be a flat distribution around the value @xmath150 in between the two maxima and minima so this slightly favours perpendicular orientation of the patches along the forward ( or backward ) direction and parallel orientation along the transversal direction . under low coverage ( @xmath130 )
conditions , on the other hand , there is clear evidence of a parallel orientation of the patches along the forward ( or backward ) direction , the opposite being true for a perpendicular orientation of the patches .
this confirms the tendency to filament formations previously alluded to .
the situation is even more evident at low density , @xmath132 , as shown in the right two panels ( b ) and ( d ) .
additional insights on the angular correlations of patch distributions can be obtained by considering other coefficients @xmath151 of the rotational invariants @xmath152 as defined in eqs .
( [ appa : eq1 ] ) and ( [ appa : eq2 ] ) ; they have proven to be of invaluable help in discriminating between parallel and antiparallel configurations occurring in different models such as dipolar hard spheres @xcite and heisenberg spin fluids . @xcite
some relevant properties of these coefficients are also listed in appendix [ app : appb ] , where we display explicit expressions for the first few coefficients . in the two - patch case
, we note that all coefficients with @xmath153 or @xmath154 odd vanish , so we depict the first nonvanishing coefficients @xmath155 , @xmath156 , @xmath157 in fig .
[ fig : fig8 ] for the same state points as before .
note that the left - most two curves , ( a ) and ( c ) , correspond to density @xmath108 , temperature @xmath109 , coverages @xmath129 ( a ) and @xmath130 ( c ) , and are plotted on the same scale .
while qualitative trends are similar , the two cases have significantly different behavior . within a given coverange @xmath155 and @xmath156 are almost coincident , with positive correlation in the well region @xmath158 , whereas @xmath159 has decreasing positive correlation in the same region and negative correlation for @xmath160 .
numerical values , on the other hand , differ among each other , with small values for @xmath155 and @xmath156 in the high coverage case @xmath129 and significantly higher values in the low coverage limit @xmath130 . likewise ,
for the low - density state point @xmath132 , @xmath109 , right - hand - side plots ( b ) and ( d ) can be unambiguously discriminated between high ( b ) and low ( d ) coverages .
we shall return to this point in the comparison with the one - patch results , where the physical meaning of these coefficients will be discussed .
figure [ fig : fig9 ] shows the critical parameters for the case of particles with one and two patches , both reported as a function of the total coverage . here only mc results are reported in view of their precision and reliability .
the @xmath29 limit corresponds to the sw case and coincides for both models .
an analogous figure , for the case of adhesive patchy spheres ( the limit of the present model for vanishing ranges ) , has been reported in ref . within a simplified integral equation scheme .
with respect to the adhesive limit , the range of coverages which can be explored numerically is significantly wider ( for both one - patch and two - patch cases ) . in the case of two patches , crystallization pre - empts the possibility of exploring the smaller values . in the one - patch case ,
the process of micelle formation , also observed experimentally , @xcite suppresses the phase - separation process @xcite at small @xmath0 values .
the critical parameters decrease on decreasing @xmath0 for both one - patch and two - patch models .
the behavior of @xmath161 can be explained on the basis of a progressive reduction of the attractive surface .
the decrease in the critical density becomes significantly pronounced only for the smallest @xmath0 values and can be attributed to the lower local density requested for extensive bonding .
such behavior is analogous to the suppression of the critical density observed when the particle valence decreases .
@xcite in the @xmath0 region where it is possible to evaluate the critical parameters , @xmath161 and @xmath162 for the two - patch case are always larger than the corresponding one - patch values , a trend which can be tentatively rationalized on the basis of the ability to form a larger number of contacts and higher local bonded densities for the case in which both poles of the particles can interact attractively with their neighbors .
no evidence of a crossing between the two geometries is observed .
such a crossing has been predicted by a theoretical approach based on a virial expansion up to third order in density and appropriate closures of the direct correlation function . @xcite
it is worth emphasizing that the above dependence on the number of patches , at a given coverage , provides clear evidence of the impossibility of rationalizing the change of the critical line on the basis of a trivial decrease of the attractive strength of interactions due to the reduction in coverage , as alluded to in section [ subsec : coex ] . for the sake of completness
, we also report rhnc integral equation results for the most relevant thermodynamic quantities , as a function of the coverage @xmath0 .
these are shown in table [ tab : tab2 ] for the same high - density state , @xmath108 at @xmath109 , considered above for structural information . here
we present the reduced internal energy per particle and the reduced excess free energy per particle , @xmath163 and @xmath164 , respectively , the reduced chemical potential @xmath165 , the compressibility factor @xmath166 , and the inverse compressibility @xmath167 .
these results may be compared with those of table iv in ref . listing the same quantities for the one - patch counterpart .
( we ignore the tiny difference in densities between the two calculated states . )
the last two columns give the reference hs diameter @xmath43 stemming from the variational rhnc scheme ( see ref . for details ) and the average coordination number within the wells @xmath168 , whose one - patch counterparts are included in tables iv and v of ref .
, respectively .
note that @xmath168 here is systematically larger than in the one - patch case , in qualitative agreement with the mc results of fig .
[ fig : fig9 ] . within the rhnc integral approach
, the analysis of the angular distribution of patches within the first shell given in section [ subsec : angular_dist2p ] revealed that the cylindrical symmetry of a pair of patches ( 2p case ) on each particle was very effective in driving the system to morphologically different configurations in the low ( @xmath169 ) and high ( @xmath170 ) coverage limits , as illustrated in fig .
[ fig : fig7 ] .
it is natural to expect a very different situation in the single - patch case ( 1p case ) .
this is indeed the case as further elaborated below . for the single patch with @xmath130 coverage ( fig .
[ fig : fig10 ] , bottom panels ( c ) and ( d ) ) parallel patches ( @xmath171 ) are more likely in the perpendicular direction ( @xmath172 or @xmath173 ) , whereas antiparallel patches ( @xmath174 ) are more likely in the forward direction ( @xmath175 ) . the case of perpendicular patches ( @xmath176 ) are conversely more or less equally distributed along the whole angular region @xmath177 .
there is no qualitative difference between the situation of high ( @xmath108 ) and low ( @xmath132 ) densities as shown by the contrast between the bottom left ( c ) and right ( d ) panels .
note that the result significantly contrasts with the corresponding results of the two - patches case depicted in fig .
[ fig : fig7 ] .
consider now the opposite situation of very large coverage ( @xmath129 ) ( fig .
[ fig : fig10 ] , ( a ) ) where there is a single well - defined peak for antiparallel orientations ( @xmath174 ) in the backward direction ( @xmath178 ) .
again , this markedly differs from the two - patch case ( fig .
[ fig : fig7 ] , top left panel ( a ) ) , where there is a triple peak for aligned patches ( @xmath179 ) in the forward ( @xmath180 ) , perpendicular ( @xmath173 ) , and backward ( @xmath178 ) orientations .
this is a dense state point . under diluted conditions , @xmath132
, we find a qualitatively similar behavior as in the dense case , with antiparallel alignment in the forward direction ( which can not be physically distingushed from the backward one ) .
clearly the predominant antiparallel alignment is reflecting the tendency to micellization rather than polymerization which is built into the single patch symmetry .
it is also interesting to contrast the coefficients of rotational invariants for the one - patch case with those obtained in the two - patch counterpart in fig .
[ fig : fig7 ] . at variance with the two - patch case , here
all coefficients are nonvanishing so that we consider the first nonvanishing instances , that is @xmath181 , @xmath182 , and @xmath155 , which are particularly useful as giving the projections over the important invariants .
@xcite we have evaluated these coefficients for the same state points considered in the two - patch case in fig . [ fig : fig11 ] for both dense or diluted conditions and small or large coverages .
in contrast with the two - patch case , here the effect of coverage appears to be less significant , as can be inferred by inspection of the dense case @xmath108 ( left panels ( a ) and ( c ) ) . for the @xmath181 case the projection coefficient along the ferroelectic invariant @xmath183 in appendix [ app : appb ] we find a negative correlation within the well both for @xmath129 ( top left panel ( a ) ) and @xmath129 ( bottom left panel ( c ) ) , as expected from the tendency to form antiparallel alignments .
likewise , the projection @xmath182 along the dipolar invariant @xmath184 is found to be negative and numerically similar to @xmath181 at both coverages , again indicating negative correlation to dipolar alignment . the only positive correlation is found for the @xmath155 component , which does not distinguish between up and down symmetry , in qualitative agreement with the two - patch analogue . this situation is replicated in the diluted case ( right two panels ( b ) and ( d ) ) with different numerical values , thus indicating that these correlations are signatures of robust orientational trends induced by the the particular one - side symmetry of the single - patch potential
we have performed a detailed study of a fluid whose particles interact via a two - patch kern - frenkel potential that attributes a negative square - well energy whenever any two patches on the spheres are within a solid angle associated with a predefined coverage and within a given distance given by the well width , and a simple hard - sphere repulsion otherwise .
this model can be reckoned as a paradigm of a unit system with incompatible elements ( _ e.g. _ , hydrophobic and hydrophilic ) that can self - assemble into different complex superstructures depending on the parameters of the original unit ( _ e.g. _ , coverage ) .
we have exploited state - of - the - art numerical simulations ( standard metropolis , gibbs ensemble , and grand canonical monte carlo ) coupled with rhnc integral equation theory following the approach outlined in previous work on a single patch .
@xcite on comparing rhnc integral equation with numerical simulations , we find the former to be quantitatively predictive in a large region of coverage , even close to the gas - liquid transition critical region , over a range of coverage which is significantly larger than the single patch counterpart .
the reason for this is attributed to the fact that rhnc uses the approximated hard - sphere bridge function , which retains spherical information , as a unique approximation throughout the entire calculation , a feature which works better for the more symmetric two - patch case than the highly asymmetric one - patch kern - frenkel potential .
having assessed the reliability of the rhnc integral - equation approach , we have fully exploited its capabilities in providing detailed angular information that is typically inaccessible to mc simulations , as already discussed in ref . .
this has been done in two ways .
first , by computing the orientational distribution probability of parallel and perpendicular alignment of patches within a spherical shell in the region @xmath135 . this methodology is able to account for the erratic coverage dependence of the pair correlation function @xmath4 by clearly discriminating between small and large coverages at all densities .
the same approach also enlightens the characteristic symmetries of the patch distributions when the two - patch case result is contrasted with the one - patch analogue .
second , by computing the rotational - invariant coefficients that are the projections of @xmath4 over rotational invariants .
again , this can discriminate between small and large coverages ( at all densities ) and single and double patches .
our monte carlo results extend those originally obtained by kern and frenkel @xcite for the two - patch case and can be contrasted with those of the corresponding single - patch counterpart @xcite and those obtained when the radial part of the potential is of the baxter type .
@xcite the rhnc calculation presented here , along with the corresponding calculation carried out in our previous paper , @xcite together constitute the first attempt to apply a well - defined integral equation theory to such highly anisotropic potentials having sharp angular modulation . an important outcome of our calculations is the clarification of the combined effect that size and distribution of the patches have on the gas - liquid coexistence lines and critical parameters .
the reduction of the bonding surface clearly decreases the critical temperature , an effect which can be related to the decrease in the bonding energy of the system .
more interestingly , it also shows a suppression of the critical density , which can be interpreted along the same lines used in interpreting the valence dependence in patchy colloids .
@xcite indeed , the maximally bonded structures require lower and lower local densities on decreasing @xmath0 . interestingly , while in the single - bond - per - patch condition the evolution of the critical parameters on decreasing valence can be followed down to the limit where clustering prevents phase separation , @xcite in the model studied here crystallization pre - empts the observation of the liquid - gas separation for @xmath185 .
crystallization is here much more effective due to the analogy between the local fully - bonded configuration and the crystal structure .
by contrast , crystallization is never observed for the one - patch case , where it has been shown that the lowest energy configuration is reached instead via the process of formation of large aggregates ( micelles and vesicles ) or via the formation of lamellar phases .
@xcite this difference highlights the important coupling between the orientational part of the potential and the possibility of forming extended fully - bonded structures .
our results indicate that , for a given coverage , in the two - patch case both the critical temperature and density are slightly higher then their corresponding one - patch counterparts , thus indicating that an increase of the valence favours the gas - liquid transition , in agreement with previous findings .
a final important consequence of our study concerns the limit of very small coverages that is particularly interesting .
indeed , it is possible to tune the structure of the system and control the topology of its ordered arrangement . by doing this
we have observed a progression from the case where chains are stable ( in the one - bond - per - patch limit ) to the case where independent planes are found , evolving for slightly larger @xmath0 values into an ordered three - dimensional crystalline structure . each of these ordered structures is observed in a restricted range of @xmath0 values .
this possibility of fine tuning the morphology by controlling the patterning of the particle surfaces may offer an interesting possibility for specific self - assembling structures .
an additional perspective of our work should be stressed .
several studies ( see _ e.g. _ ref . and
references therein ) have exploited spherically - symmetric potentials to mimic effective protein - protein interactions , especially in connection with protein crystallization .
@xcite this is clearly unrealistic for the majority of proteins where the distribution of hydrophobic surface groups is significantly irregular , a feature that can be captured , at the simplest level of description , by the model studied here .
the specific location of the coexistence lines , such as those considered in the present study , have important consequences in the study of pathogenic events for sickle cell anemia @xcite and other human diseases . @xcite
we thank philip j. camp and enrique lomba for useful suggestions .
fs acknowledges support from noe softcomp nmp3-ct-2004 - 502235 , erc226207patchycolloids and itn - comploids .
ag acknowledges support from prin - cofin 2007b57eab(2008/2009 ) .
the expansion in spherical harmonics @xmath186 of @xmath4 in an arbitrary space frame reads @xcite @xmath187 where we have introduced the rotational invariants @xcite @xmath188 note that @xmath189 coincides with @xmath190 up to a normalization constant ( see appendix [ app : appb ] ) .
we are free to set the origin of the coordinate frame at the center of particle @xmath8 and choose its orientation so that @xmath191 without loss of generality .
we first note that @xmath192 clearly , the orientation of the @xmath193 axes is then irrelevant , so we may integrate out the angles @xmath194 and @xmath143 ; this leads to the average @xmath195 , where we note that @xmath196^{1/2 } p_{l_{2 } m_{2}}\left(\cos \theta_2\right ) \frac{1}{2\pi } \int_{0}^{2\pi } d\varphi_2 e^{\mathrm{i } m_2 \varphi_2 } \\
\nonumber & = & \left ( \frac{2l_2 + 1}{4\pi } \right)^{1/2 } p_{l_{2}}\left(\cos \theta_2\right ) \delta_{m_20}\end{aligned}\ ] ] and similarly for @xmath197 .
here the @xmath198 are the usual legendre polynomials .
we have then from eqs .
( [ appa : eq1 ] ) and ( [ appa : eq2 ] ) that @xmath199^{1/2 } c\left(l_{1 } l_{2 } l ; 0 0 0 \right ) p_{l_{2}}\left(\cos \theta_2\right ) p_{l}\left(\cos \theta\right ) .
\label{appa : eq5}\end{aligned}\ ] ] we are interested in the angular behavior within the well , @xmath131 , so we finally integrate over the radial variable @xmath66 and define @xmath200 the result is then a function of the polar coordinate @xmath144 of @xmath10 and the polar orientation of the second patch , @xmath146 ; that is , @xmath201^{1/2 } c\left(l_{1 } l_{2 } l ; 0 0 0 \right ) p_{l_{2}}\left(\cos \theta_2\right ) p_{l}\left(\cos \theta\right),\end{aligned}\ ] ] given that the @xmath202 axis is aligned with the patches of particle 1 .
in this appendix we consider the coefficients @xmath190 of rotational invariants that have proven to be particularly useful in discriminating among different orientational behaviors . in numerical simulations
they are defined as follows ( see _ e.g. _ ref . ) , @xmath204 where the @xmath205 are rotational invariants .
explicit expressions for the first few are @xcite @xmath206 , \\
\nonumber \delta^{220}\left(12 \right)&= & \frac{5}{2 } e\left(12\right)=\frac{5}{2 } \left[3\left ( \hat{\mathbf{n}}_1 \cdot \hat{\mathbf{n}}_2\right)^2 -1\right].\end{aligned}\ ] ] other expressions can be found in ref . .
we note that the first expression in eqs .
( [ appb : eq2 ] ) yields @xmath207 , which coincides with the radial distribution function @xmath208 .
here we have @xmath209 ; in all other cases @xmath210 .
some of the coefficients have particularly interesting physical interpretations : the term @xmath181 is the coefficient of ferroelectric correlation , the term @xmath182 the coefficient of dipolar correlation , the term @xmath155 the coefficient of nematic correlation , and so on .
it might be useful to show how these general expressions ( typically computed in monte carlo simulations ) connect with the corresponding ones typically evaluated in an integral equation approach .
we do this for the representative case of @xmath211 , the others being similar .
we define @xmath212 , where the proportionality constant is obtained through a particular prescription to be further elaborated below .
the projections @xmath213 of @xmath4 on the rotational invariants @xmath214 as defined by eq .
( [ appa : eq1 ] ) are related to the values @xmath215 in the axial frame by @xcite @xmath216 where @xmath217 .
consider as a representative example the quantity @xmath218 defined in eqs .
( [ appb : eq2 ] ) and note that in eq .
( [ appb : eq1 ] ) one has @xmath219 upon choosing the molecular frame ( @xmath220 ) , one finds @xmath221,\end{aligned}\ ] ] where the @xmath186 are spherical harmonics . @xcite
the expansion ( [ appa : eq1 ] ) in a molecular frame reduces to @xmath222 using eqs .
( [ appb : eq5 ] ) , ( [ appb : eq6 ] ) , and the orthogonality relations for spherical harmonics , @xcite one finds easily that @xmath223,\end{aligned}\ ] ] so that , combining with @xmath182 given in eqs .
( [ appb : eq1 ] ) and ( [ appb : eq2 ] ) and using the symmetry property @xmath224 , one finds @xmath225 we can work out the first few projections by using tabulated values for the clebsch - gordan coefficients @xmath226 . using the symmetry properties of the clebsch - gordan coefficients and the @xmath227
, one finds @xmath228 , \\
\nonumber g\left(r;112\right)&= & \left ( 8\pi/15 \right)^{1/2 } \left [ g_{110}\left(r\right ) + g_{111}\left(r\right ) \right ] , \\
\nonumber g\left(r;220\right)&= & \left ( 4\pi/5 \right)^{1/2 } \left[g_{220}\left(r\right)-2
g_{221}\left(r\right)+2g_{222}\left(r\right ) \right ] , \\
\nonumber g\left(r;222\right)&= & -\left ( 8\pi/35 \right)^{1/2}\left[g_{220}\left(r\right ) -g_{221}\left(r\right)-2 g_{222}\left(r\right ) \right ] , \\
\nonumber g\left(r;224\right)&= & \left ( 8\pi/35 \right)^{1/2 } \left[g_{220}\left(r\right)+\frac{4}{3 } g_{221}\left(r\right)+ \frac{1}{3}g_{222}\left(r\right ) \right ] , \\
\nonumber g\left(r;011\right)&= & \left ( 4\pi/3 \right)^{1/2 } g_{010}\left(r\right ) = -g\left(r;101\right ) , \\
\nonumber g\left(r;022\right)&= & \left ( 4\pi/5 \right)^{1/2 } g_{020}\left(r\right)=g\left(r;202\right ) , \\
\nonumber g\left(r;121\right)&= & -\left ( 8\pi/15 \right)^{1/2 } \left[g_{120}\left(r\right)-\sqrt{3}\ , g_{121}\left(r\right ) \right]=-g\left(r;211\right ) , \\
\nonumber g\left(r;123\right)&= & \left ( 12\pi/35 \right)^{1/2}\left[g_{120}\left(r\right)+ \frac{2}{\sqrt{3 } } g_{121}\left(r\right ) \right]=-g\left(r;213\right ) .
\label{appb : eq9}\end{aligned}\ ] ] as anticipated above , we now fix the proportionality constant in @xmath212 by dividing out the leading constants above so that the coefficient of @xmath229 is unity ; _ e.g. _ , @xmath230 .
0.5 cm and @xmath235 ( @xmath236 ) , where the orientation vectors @xmath141 , @xmath142 define angles @xmath237 , @xmath238 with the vector @xmath10 joining the centers of the two spheres and directed from sphere @xmath8 to sphere @xmath9 .
the particular case shown corresponds to a @xmath239 fraction of attractive surface ( coverage @xmath0 ) .
[ fig : fig1],title="fig:",width=529 ] 0.5 cm .
points represent mc results while thick solid lines report rhnc values ( for @xmath240 ) .
thinner solid lines are a guide for the eye whereas the thick dashed line shows the estimated gcmc critical point for any fixed coverage @xmath0 .
the mc data for @xmath241 ( sw ) coincide within the numerical error with the ones reported in refs . and .
[ fig : fig2],title="fig:",width=529 ] 0.5 cm . ( a )
the case of coverages such that each patch can be involved in only one interaction . in this case , the system forms polydisperse chains .
the snapshot here refers to the case @xmath242 and @xmath243 .
( b ) values of @xmath0 such that each patch can be involved in only two interactions . in this case , at low @xmath80 the system forms bonded planes interacting with each other only via excluded volume interactions .
the snapshot shows one such plane .
( c ) values of @xmath0 such that each patch can be involved in only three interactions . the crystal is now formed by interconnected planes , with a triangular arrangement of the particles in the plane .
adjacent planes are shifted in such a way that each particle sits in correspondence to the center of a triangle of the previous and following planes .
( d ) values of @xmath0 such that each patch can be involved in at most four interactions . the crystal is now formed by interconnected planes , with a square arrangement of the particles in the plane .
adjacent planes are shifted in such a way that each particle sits in correspondence to the center of a square of the previous and following planes .
[ fig : fig3],title="fig:",width=529 ] from the rhnc equation for different coverages and two specific orientations of the patches : @xmath110 configuration corresponding to @xmath106 ( a ) and x configuration corresponding to @xmath107 ( b ) .
all curves are for a state point with @xmath71 , @xmath108 , and @xmath109 .
black and green lines show the limiting cases of square - well ( @xmath29 ) and hard - sphere ( @xmath30 ) potentials , respectively . other coverages
are @xmath244 for both the @xmath110 and x configurations .
[ fig : fig4],width=453 ] for different coverages and two specific orientations of the patches : @xmath110 configuration corresponding to @xmath106 ( a ) and x configuration corresponding to @xmath107 ( b ) .
all curves are for a state point with @xmath71 , @xmath108 , and @xmath109 .
both rhnc and mc results are depicted for @xmath245 . [
fig : fig5],width=453 ] ( ( a ) and ( c ) ) and the radial distribution function @xmath126 ( ( b ) and ( d ) ) for coverages @xmath129 and @xmath130 , respectively . here
@xmath109 and @xmath71 as before , while densities are @xmath108 and @xmath132 .
[ fig : fig6],width=529 ] as functions of @xmath246 for two different orientations of the patches on sphere @xmath9 , given that sphere @xmath8 is fixed with patches along the @xmath147 axis .
results are reported for two different coverages , @xmath129 ( ( a ) and ( b ) ) and @xmath130 ( ( c ) and ( d ) ) , and two different densities , @xmath108 ( ( a ) and ( c ) ) and @xmath132 ( ( b ) and ( d ) ) , at the same temperature @xmath109 .
again , the square - well width was set to @xmath71 .
the colored arrows are cartoons of the orientations of sphere @xmath9 patches , corresponding to @xmath247 .
note that these are the same state points considered in fig .
[ fig : fig6 ] .
[ fig : fig7],width=529 ] , @xmath159 , and @xmath156 as functions of @xmath138 .
results are reported for two different coverages , @xmath129 ( ( a ) and ( b ) ) and @xmath130 ( ( c ) and ( d ) ) , and two different densities , @xmath108 ( ( a ) and ( c ) ) and @xmath132 ( ( b ) and ( d ) ) , at the same temperature @xmath109 and square - well width @xmath71 .
again , these are the same state points considered in figs . [
fig : fig6 ] and [ fig : fig7 ] .
[ fig : fig8],width=529 ] for the one - patch model as a function of @xmath246 for three different orientations of the patch on sphere @xmath9 , given that sphere @xmath8 is fixed with patch along the @xmath147 axis .
this is the one - patch counterpart of fig .
[ fig : fig7 ] . results are reported for two different coverages , @xmath129 ( ( a ) and ( b ) ) and @xmath130 ( ( c ) and ( d ) ) , and two different densities , @xmath108 ( ( a ) and ( c ) ) and @xmath132 ( ( b ) and ( d ) ) , at the same temperature @xmath109 and square - well width @xmath71 .
the colored arrows are cartoons of the orientations of the sphere @xmath9 patch , corresponding to @xmath248 .
[ fig : fig10],width=529 ] , @xmath182 , and @xmath155 for the one - patch model as functions of @xmath138 .
results are reported for two different coverages , @xmath129 ( ( a ) and ( b ) ) and @xmath130 ( ( c ) and ( d ) ) , and two different densities , @xmath108 ( ( a ) and ( c ) ) and @xmath132 ( ( b ) and ( d ) ) , at the same temperature @xmath109 and square - well width @xmath71 .
again , these are the same state points considered in fig . [
fig : fig8 ] .
[ fig : fig11],width=529 ] | we report on a computer simulation and integral equation study of a simple model of patchy spheres , each of whose surfaces is decorated with two opposite attractive caps , as a function of the fraction @xmath0 of covered attractive surface .
the simple model explored the two - patch kern - frenkel model interpolates between a square - well and a hard - sphere potential on changing the coverage @xmath0 . we show that integral equation theory provides quantitative predictions in the entire explored region of temperatures and densities from the square - well limit @xmath1 down to @xmath2 . for smaller @xmath0 , good numerical convergence of the equations
is achieved only at temperatures larger than the gas - liquid critical point , where however integral equation theory provides a complete description of the angular dependence .
these results are contrasted with those for the one - patch case .
we investigate the remaining region of coverage via numerical simulation and show how the gas - liquid critical point moves to smaller densities and temperatures on decreasing @xmath0 .
below @xmath3 , crystallization prevents the possibility of observing the evolution of the line of critical points , providing the angular analog of the disappearance of the liquid as an equilibrium phase on decreasing the range for spherical potentials .
finally , we show that the stable ordered phase evolves on decreasing @xmath0 from a three - dimensional crystal of interconnected planes to a two - dimensional independent - planes structure to a one - dimensional fluid of chains when the one - bond - per - patch limit is eventually reached . |
the last decade has seen the first instances of statistical studies with quasars using purely photometric samples .
examples of these include the measurement of the integrated sachs
wolfe effect @xcite and cosmic magnification bias @xcite , and studies of the clustering of quasars on large @xcite and small @xcite scales .
the importance of photometrically classified quasar samples will only increase during the next decade as large new imaging surveys will uncover large samples of quasars at fainter magnitudes , with minimal spectroscopy for the faintest objects .
while efficient photometric classification is one requirement to facilitate studies of quasars without extensive spectroscopy , it has also been crucial to develop accurate methods for quasar redshift estimation based on broadband photometry .
techniques for photometric redshift estimation have long been successful for galaxies ( e.g. , * ? ? ?
* ; * ? ? ?
* ) and became feasible for quasars with the advent of precise multi - filter photometry @xcite .
closely related to the quasar photometric - redshift problem traditionally seen as a _ regression _ problem is the question as to how best to perform photometric _ classification _ of quasars .
it has become clear that the best classifiers are probabilistic in nature in that they calculate probabilities for objects to be quasars based on accurately calibrated models for stellar and quasar photometry ( e.g. , * ? ? ?
* ; * ? ? ?
these probabilities are often calculated for quasars in certain broad redshift ranges , and they therefore also act as low - resolution photometric redshifts for the objects they classify as quasars . the object classification technique of @xcite uses bins of width @xmath1 and , thus , achieves classification with a finer photometric - redshift estimate . however , detailed photometric redshift estimates for photometrically classified quasars utilize heterogeneous techniques , such that the resulting redshift probability distributions are inconsistent with the broad probabilities used for the initial quasar classification . for instance , this is the case for the photometric quasar catalogs of @xcite . for these catalogs , a non - parametric kernel - density - estimation ( kde ) technique that ignores photometric uncertainties
was used to classify quasars , while a parametric model that convolves the quasar color locus with the photometric uncertainties a single gaussian distribution in bins of redshift @xmath2was applied to estimate redshift @xcite . for many purposes
, one would like to target quasars in arbitrary redshift ranges that differ from those predetermined and imposed by a broad classification method .
for example , the baryon oscillation spectroscopic survey ( _ boss _ ; @xcite ) of _ the sloan digital sky survey iii _ ( _ sdss - iii _ ) aims to measure the baryon acoustic feature in the ly@xmath3 forest of medium - redshift ( @xmath4 ) quasars ( e.g. , * ? ? ?
* ; * ? ? ?
the spectral range accessible to the _ boss _ spectrographs is @xmath5 ( barkhouser et al . , 2011 , in preparation ) , thus _ boss _ can only study the ly@xmath3 forest as traced by redshift @xmath6 quasars .
therefore , _ boss
_ requires quasars to be targeted based on their probability to be at redshift @xmath7 , and the _ boss _ quasar classifiers were trained with this constraint ( e.g. , * ? ? ?
* ; * ? ? ?
however , other ground - based instruments can observe at shorter wavelengths , e.g. , the multi - object double spectrograph for the large binocular telescope , which can observe the spectral range @xmath8 @xcite .
this instrument could study the ly@xmath3 forest starting at redshift @xmath9 .
an ly@xmath3 forest experiment designed for the large binocular telescope might therefore target quasars in the redshift range @xmath10 in addition to those at higher redshift .
another example of a project that requires accurate photometric characterization of quasars is the search for binary quasars @xcite , where the key metric is the probability that two objects are both quasars _ and _ proximate in redshift , i.e. , the joint ( or `` overlapping '' ) probability that both components of a pair of objects are quasars of a particular redshift .
similarly , the search for projected quasar pairs for absorption line studies @xcite requires the joint probability that both objects in a projected pair are quasars .
such calculations require a full model of quasar probabilities and redshifts .
ideally , therefore , photometric redshift estimation and quasar classification ought to be performed together . in the specific case of objects observed using the @xmath0 filter system @xcite , quasar photometric redshifts are plagued by a host of degeneracies at redshifts where various quasar emission lines are mistaken for the ly@xmath3 line @xcite
; this results in `` catastrophic '' redshift failures , although consideration of the full redshift posterior distribution function ( pdf ) shows that most of these failures have a significant integrated probability around the correct redshift ( e.g. , @xcite ; see below ) .
the addition of non-@xmath0 data , e.g. , ultraviolet ( uv ) and near - infrared ( nir ) measurements , can both alleviate these redshift degeneracies _ and _ improve quasar star separation .
quasar classification and characterization in the infrared has been considered for simulated objects and for quasar samples with a range of depths and areas ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
the nir is also the region to search for the highest redshift quasars ( redshift @xmath11 ; * ? ? ?
these studies show the great promise that nir data hold for quasar selection and redshift estimation .
the uv holds a similar potential ( see , e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
the technique we introduce in this article , which we denote _ xdqsoz _ , is the first that deals with the simultaneous classification of quasars and assignation of quasar redshifts .
this technique extends the _ xdqso _ quasar classification technique of @xcite to model the density of quasars in color redshift space with a flexible semi - parametric model consisting of a large set of gaussian component distributions .
this model can be integrated analytically over any redshift range to calculate probabilities from flux measurements for individual objects .
this , in turn , allows quasar probabilities to be calculated over any redshift range .
thus , a probability distribution in redshift space ( a `` pdf '' ) is a natural component of the model . because we use the _ extreme deconvolution _ ( xd )
technique @xcite as our density estimation tool , the method can be trained on and applied to low signal - to - noise ratio data , even with missing values , e.g. , to objects missing measurements in any arbitrary collection of filters .
this feature allows us to naturally include uv and nir broadband fluxes , where sky coverages differ , as part of our model space and to distinguish sources that are missing data in a particular band from objects that are dropping out of that band .
we show that the addition of uv and nir broadband fluxes improves quasar star separation significantly and that it essentially breaks all of the redshift degeneracies inherent to the @xmath0 filter set .
this article is organized as follows . in section [ sec : general ] , we discuss general aspects of photometric redshift estimation and classification in the context of quasars . we briefly describe the data used to train and test the new method in section [ sec : data ] .
section [ sec : xdqsoz ] contains a full description of the _ xdqsoz _ quasar model and section [ sec : targeting ] shows how this model is used to calculate quasar probabilities over arbitrary redshift ranges .
section [ sec : photoz ] assesses the performance of the photometric redshifts obtained using the _ xdqsoz _ model .
a discussion of various extensions of the model is given in section [ sec : discussion ] and we conclude in section [ sec : conclusion ] .
the describes the photometric classification and redshift estimation _ xdqsoz _
code that is made publicly available . in what follows , ab magnitudes
@xcite are used throughout . where dereddened fluxes and magnitudes are required we have used the reddening maps of @xcite . all magnitudes and fluxes should be considered as dereddened unless mentioned otherwise .
a technique for photometric redshift estimation of quasars should have the following properties .
* it should provide full probability distributions for the redshift of the quasar based on its observed photometry , because this information has particular utility for quasars ( e.g. , * ? ? ? * ) as near - degeneracies in redshift estimation from broadband photometry are ubiquitous for quasars ( e.g. , * ? ? ?
* ; * ? ? ?
* ) . * upon the evaluation of the probability of the redshift
the photometric uncertainties should be treated properly to allow photometric redshift estimation for faint objects . *
if based on an empirical training set , the technique should be able to be trained on low signal - to - noise ratio data with potentially missing data .
the training set and the evaluation set should also be allowed to have different noise properties , e.g. , different distributions of signal - to - noise ratio .
for example , while the optical fluxes are mostly well measured for a spectroscopic training sample , the addition of uv and nir data can help break redshift degeneracies ( see below ) , but these measurements often have low signal - to - noise ratio , even for the training set .
* the technique should allow an explicit redshift prior to be specified .
the key to photometric classification and redshift estimation for quasars based on broadband fluxes is the joint probability of an object s fluxes , its redshift , and the proposition that it is a quasar @xmath12
. this joint probability can be re - written in several ways that correspond to different ways of approaching the problem @xmath13 photometric redshift estimation corresponds to the probability of an object s redshift conditioned on its fluxes and assuming that it is a quasar : @xmath14 quasar classification is the probability that an object is a quasar based on its fluxes . to classify quasars in a certain redshift range @xmath15
, we integrate the joint probability that the object is a quasar with redshift @xmath16 over redshift : @xmath17 the probability that an object is a quasar of any redshift is obtained by setting the redshift range @xmath18 $ ] .
the normalization factor @xmath19 in this equation is given by @xmath20 the probability of an object not being a quasar can be obtained empirically by modeling the fluxes of non - quasars ( see * ? ? ?
* ; * ? ? ?
the discussion above suggests that a unified approach to classification and photometric redshift estimation is possible . because the method described in this article is the first technique in this class , we briefly discuss previous attempts at photometric redshift estimation and how they fit in the framework outlined in this section .
the @xmath21-nearest neighbors approach of @xcite is an instance - based machine - learning technique that compares the colors of test objects to the @xmath21 nearest objects in color - space in a training set , and assigns a weighted combination of the redshifts of those nearest neighbors to the test object .
its generalization to take observational flux - uncertainties into account involves perturbing both the test and the training data within their gaussian noise ellipsoids @xcite . in its noiseless implementation
the method does not return a full probability distribution for the redshift . when taking the photometric uncertainties into account the technique essentially returns samples from @xmath22 as in equation ( [ eq : jointprob3 ] ) , which can be binned to obtain the full posterior distribution function .
while the photometric uncertainties of the test objects are handled correctly , the approach for dealing with the uncertainties of the training data effectively convolves with the uncertainties twice , as it adds scatter to the training data that are already scattered from the intrinsic distribution due to photometric noise .
because the technique directly uses the training set , it also implicitly applies a redshift prior that approaches the _ observed _ redshift distribution .
this choice of prior does not reflect the _ intrinsic _ redshift distribution , as the _ observed _ distribution is shaped by various selection effects @xcite . the approach taken by @xcite consists of fitting the relative - flux redshift distribution and its scatter to produce the likelihood of the quasar redshift as in equation ( [ eq : jointprob1 ] ) .
this fit is conducted without taking the flux uncertainties into account , but upon evaluation of test objects the flux uncertainties are fully handled .
closest to the approach taken in this article is the technique of @xcite .
the distribution of colors in a set of narrow bins in redshift is fit as a single multi - variate gaussian distribution .
this approach is similar to quasar classification approaches where the color or relative - flux distributions of quasars are fit in much broader redshift ranges using more general density models ( @xcite , @xcite ) .
@xcite do not use the photometric uncertainties of the data they use for training .
but , as in @xcite , photometric uncertainties for test objects are fully taken into account . all of the techniques described above could be extended to allow quasar classification by specifying the necessary factors of @xmath23 or @xmath24 in equations ( [ eq : jointprob1])([eq : jointprob3 ] ) .
the latter could be taken from a quasar classification scheme such as nbc - kde @xcite or _ xdqso _
@xcite , although care should be taken that the classification method uses the same redshift prior as the photometric redshift technique for consistency .
the _ xdqsoz _ technique introduced in this article uses equation ( [ eq : jointprob2 ] ) as the basis of both quasar classification and photometric redshift estimation .
specifically , we model the relative - flux
redshift distribution using a large number of gaussians by deconvolving this distribution for a training set using the xd technique @xcite .
we use empirical relative fluxes that are re - weighted using an explicit , magnitude - dependent redshift prior ( which can easily be divided out ) . as described in section [ sec : targeting ] , conditioning on the fluxes to obtain full photometric probability distributions for the redshift , and marginalization over redshift to classify quasars , is simple and fast in this approach . because we deconvolve the relative - flux
redshift distribution when training , we can straightforwardly incorporate uv and nir data , both of which significantly improve the accuracy and precision of the inferred redshifts .
the sloan digital sky survey ( _ sdss _ ; @xcite ) has obtained _ u , g , r , i _ and _ z _ ccd imaging of @xmath25 10@xmath26 deg@xmath27 of the northern and southern galactic sky @xcite .
_ sdss - iii _
@xcite has extended this area by approximately 2,500 deg@xmath27 in the southern galactic cap @xcite .
all the data processing , including astrometry @xcite , source identification , deblending and photometry @xcite , and calibration @xcite are performed with automated _ sdss _ software .
dr7 imaging observations were obtained over the period 2000 march to 2007 july .
the _ sdss _ training data used here are essentially the same as the data used to train the _ xdqso
_ method ; they are described in detail in @xcite .
we use a sample of 103,601 spectroscopically - confirmed redshift @xmath28 quasars from the sdss dr7 quasar catalog @xcite .
we use _ all _ of these quasars to essentially model the color redshift relation for quasars ( but see below for the detailed description of our method ) .
we combine the color
redshift relation with an apparent - magnitude dependent redshift prior obtained by integrating a model for the quasar luminosity function over the apparent - magnitude range of interest @xcite .
this prior for a few bins in apparent magnitude is shown in [ fig : redshiftprior ] ; also shown is the difference between the @xcite redshift prior and the prior derived from the @xcite luminosity function . as the sample of quasars from the _ sdss _
dr7 quasar catalog spans a wide range in luminosity that we apply to a narrow range in apparent magnitude and that we extrapolate to largely unexplored faint flux levels , we are ignoring correlations between quasar spectral properties and luminosity ( e.g. , * ? ? ? * ; * ? ? ? * ) .
these correlations mostly affect emission line shapes , such that they are washed out in broadband colors , especially compared to the intrinsic color - scatter .
in addition to @xmath0 optical data , we use uv data obtained by the _ galaxy evolution explorer _ space mission ( _ galex _ ; * ? ? ?
_ galex _
has performed an all - sky imaging survey in two uv bands ( fuv : 1350 to 1750 ; nuv : 1750 to 2750 ) down to @xmath29 and a medium - deep imaging survey that reaches @xmath30 ( e.g. , * ? ? ? * ) .
some of the data used below to test the technique described in this article comes from the medium - deep survey , while much of the data used to _ train _ our technique comes from the shallower all - sky survey ( because our training sample of quasars is drawn from the full @xmath31 deg@xmath27 _ sdss _ footprint ) , but this difference is largely offset by the fact that our training set is brighter than the faint part of the test set .
_ galex _
gr5 observations were obtained between april 2003 and february 2009 . rather than using _
catalog products @xcite we use measurements of the uv fluxes obtained by force - photometering _ galex _ images ( from _ galex _ data release 5 ) at the _ sdss _ centroids @xcite , such that we obtain low signal - to - noise psf fluxes of objects not detected by _
galex_. as we show below , these low signal - to - noise ratio observations are essential for better classification of redshift @xmath32 quasars .
we expect these measurements to be released as part of _ sdss _ data release 9 , scheduled for 2012 .
the top panel of [ fig : galexukidsssnr ] shows the distribution of signal - to - noise ratio for _ sdss _ quasars in the _ galex _ footprint .
a total of 62,661 objects lie in the _ galex _ fuv footprint , 63,372 lie in the nuv footprint , and 62,628 are covered by both bandpasses .
we also use nir data to improve quasar classification and photometric redshift estimation . the _ ukirt infrared deep sky survey _ ( _ ukidss _ )
is defined in @xcite and consists of five survey components with different wavebands , depths and footprints . for the study in this paper we use data from the _ ukidss _ _ large area survey _ ( _ las _ ) .
technical details about the _ ukidss _ _ las _ observing strategy are described in @xcite .
the _ ukidss _ _ las _ aims to cover 4,000 deg@xmath27 of the sdss footprint in the y , j , h and k wavebands . in this paper
we use data from _ ukidss _ _ las _ dr7 which includes observations obtained between may 2005 and july 2009 inclusive .
the _ ukidss _
_ dr7 overlaps the _
_ imaging footprint over @xmath25 2500 deg@xmath27 and has median point source 5-sigma ab magnitude limits in y , j , h and k of 20.9 , 20.6 , 20.2 , and 20.2 , respectively .
the _ ukidss _ data are acquired with the ukirt wide field camera ( wfcam ; @xcite ) . the _ ukidss _
photometric system is described in @xcite , and the calibration is described in @xcite . the pipeline processing and science archive are described in m. j. irwin et al .
( 2012 , in preparation ) and @xcite . as in the case of the _ galex _ data described in section [ sec : galexdata ]
, we use force - photometered nir fluxes at _ sdss _ positions rather than _ ukidss _ catalog data .
this `` list - driven '' information is derived from aperture photometry on data release 7 of the _ ukidss _ _ las_. we choose an aperture radius of 1 arcsec . of the 103,601 quasars in the _ sdss _
dr7 quasar sample , 29,726 lie within the _ ukidss _
dr7 k - band footprint .
approximately 22,000 of these quasars are detected in the k band with also overlapping coverage in all four observed wavebands ( y , j , h and k ) in the _ ukidss _ _ las _
dr7 source catalog .
the bottom panel of [ fig : galexukidsssnr ] shows the distribution of signal - to - noise ratio in the nir for quasars in our training sample . unlike for _ sdss
_ imaging , measurements in all four filters of the _ ukidss _ survey are not obtained during the same observing run
h and k observations are performed in the same observing block , while y and j are conducted separately .
thus the sky coverage in different _ ukidss _ bands varies , and many objects are missing data in one or more of the four bands .
the breakdown of training quasars with observations in the nir by bandpass is : y : 26,876 ; j : 27,328 ; h : 28,911 ; k:29,726 .
a total of 25,510 objects have measurements in all four bandpasses .
the differing epochs of the _ ukidss _ , _ galex _ , and _ sdss _ can range up to 9 years in the observed frame . for a quasar with a redshift of 2
this is 3 years in the rest frame . the observed optical variability in radio quiet quasars over the rest frame 2 to 5 year timescale
is observed to be in the range 0.1 - 0.2 mag and is a function of absolute magnitude @xcite .
@xcite find that the variability amplitude decreases with rest - frame wavelength by a factor of two between 1500 and 6000 with an amplitude of @xmath25 0.15 mag at 6000 ( see also @xcite ) .
@xcite have studied the mid-_ir _ variability using multi - epoch spitzer observations of a sample of @xmath25 1000 active galactic nuclei and find that the rest - frame j band variability amplitude in the rest - frame timescale is @xmath25 0.1 mag . in summary
the quasars in this study are expected to vary by @xmath25 0.3 and 0.1 magnitudes in the uv and nir , respectively , over the elapsed period of the observations and this is less than the average photometric errors in the individual wavebands and significantly less than the range in colors .
the photometric redshift technique _ xdqsoz _ is an adaptation of the _ xdqso _ technique @xcite to include redshift explicitly in the model for the quasar population .
_ achieves this by modeling the @xmath33 factor in equation ( [ eq : jointprob2 ] ) , where the _ xdqso _ technique modeled @xmath34 in three bins in redshift ( corresponding to low- , medium- , and high - redshift quasars ) . as discussed in section [ sec : general ] , this approach allows us to obtain full posterior distribution functions for the redshift of a photometrically classified quasar based on its broadband fluxes .
integrating this redshift probability distribution over a range of redshifts and properly normalizing this result using equation ( [ eq : normalization ] ) gives a photometric quasar probability in the chosen redshift range that is , as we show below , competitive with the best available photometric quasar classification techniques , e.g. , _
xdqso_. to estimate the density of quasars in flux redshift space we use _ _ extreme deconvolution _ _ @xcite . as described in section [ sec : data ] , our training set consists of the _ sdss _ dr7 quasar sample , which consists mostly of bright , viz . , dereddened @xmath35 mag ( @xmath36 mag for @xmath37 sources ) , objects with small photometric uncertainties .
the _ galex _ and _ ukidss _ data described in sections [ sec : galexdata ] and [ sec : ukidssdata ] are much shallower than the _ sdss _
data and many objects are not detected at high significance in these surveys , such that photometric uncertainties are not insignificant ( see [ fig : galexukidsssnr ] ) .
additional complications are that these two supplemental surveys have not observed the full _ sdss _ footprint and that the _ ukidss _ _ las _ footprint is different for the different nir bands , such that we have heterogeneous missing data and heteroscedastic uncertainties .
xd is uniquely suited to deal with these complications in the proper probabilistic manner .
xd assumes that the flux uncertainties are known and that they are close to gaussian , as is the case for psf fluxes for point - sources in _ sdss
_ @xcite .
we assume that the spectroscopic redshifts have vanishing uncertainties because their typical value of @xmath38 @xcite is orders of magnitude smaller than typical uncertainties in broadband photometric redshifts , which are set by the width of the quasar locus .
xd models the underlying , deconvolved distribution as a sum of @xmath39 @xmath40-dimensional gaussian distributions , where @xmath39 is a free parameter that is set using an external objective ( see section [ sec : modelconstruction ] ) .
xd consists of a fast and robust algorithm to estimate the best - fit parameters of the gaussian mixture .
the full quasar - density model is constructed by fitting the flux
redshift density of quasars in a number of bins in the @xmath41-band magnitude .
as we use the same set of quasars in each bin with a different redshift prior see the discussion in section [ sec : data]we could instead have fit a single bin , e.g. , the brightest .
the other bins could have been constructed by dividing out the redshift prior of the first bin and multiplying in the redshift priors for the fainter bins . however , as we will show below , the advantage of a gaussian representation of the flux
redshift density is that it allows integrals of this density over arbitrary redshift ranges to be calculated analytically .
this leads to fast quasar - probability estimation .
if we were instead to divide out the redshift prior and multiply in a different redshift prior , the resulting function would no longer be gaussian and the numerical integration over redshift would be much more computationally expensive . because our short - term objective is to run this algorithm on essentially all of the @xmath42 _ sdss _ point sources and in the future on the @xmath43 pb of _ lsst _ catalog data @xcite , this computational advantage is important . after fitting the first bin ,
all other fits are initialized using the previous bin s optimal solution ; these extra fits all converge very quickly as the quasar flux
redshift density does not vary strongly with apparent magnitude .
the redshift prior is shown for a few bins in apparent magnitude in [ fig : redshiftprior ] .
if a different redshift prior is desired , one can divide out this prior and multiply in a new prior ( these priors are included in the code release described in the ) . for example , if one would prefer to use the @xcite model for the quasar luminosity function , one would multiply the posterior distribution function for the redshift obtained using the fiducial @xcite prior with the factor shown in the bottom panel of [ fig : redshiftprior ] .
this panel shows the ratio of the @xcite redshift prior to the @xcite prior in a number of apparent - magnitude bin .
it is clear that there is only a significant difference at relatively large redshift and at faint magnitudes , where constraints on the luminosity function are sparse . as for the _ xdqso _ technique ,
we divide the quasar - density model into a factor describing , essentially , the color
redshift density of quasars
but we again use relative fluxes rather than colors and another factor describing the apparent - magnitude distribution of quasars .
we adopted this approach because the flux density of quasars has a dominant power - law shape corresponding to the number counts as a function of apparent magnitude , while the color distribution is much flatter .
we write @xmath44 the apparent - magnitude factor does not depend on redshift .
so , this factor is the same as used in the _ xdqso _ method .
the factor is calculated by the sum of the apparent - magnitude priors in 1 of @xcite weighted by the quasar densities in 1 of the same article . as quasar redshifts are always positive , we model the logarithm of the redshift .
since our training sample consists of point - like objects at redshift @xmath45 , about all at @xmath46 , our model should only be trusted to return reasonable densities within this range . in each bin we model the @xmath40-dimensional relative - flux
redshift density , where @xmath40 is the number of independent colors plus one ( for redshift ) , of quasars using 60 gaussians that are allowed to have arbitrary means , variance matrices , and amplitudes the amplitudes are constrained to sum to one .
we use the full set of 103,601 , @xmath47 quasars to train the final model , but in order to test whether we are under- or overfitting the data we performed a cross - validation test . to cross -
validate we extract a random subset containing 10percent of the full sample to use as an independent test data set . by training the model on the remaining 90percent of the sample we can select the number of gaussians that optimally predicts i.e .
, predicts with the highest probability the redshifts of objects in the test sample .
the results from this procedure are shown in [ fig : testloglike ] . our ability to better predict the test redshifts saturates around @xmath48
; we chose 60 gaussians to represent the relative - flux density of quasars .
compared to the _ xdqso _ method , which used 20 gaussians each in three redshift bins , this revised approach uses the same number of gaussians while representing an extra dimension ( redshift ) .
one might be concerned that because the gaussians will preferentially be found in high - density , viz . , low - redshift , regions , the density of medium- and high - redshift quasars is not adequately described in the _ xdqsoz _ model .
we will see below that this is not the case and that _ xdqsoz _ performs similarly to _ xdqso _ in selecting medium- and high - redshift quasars .
the full model consists of 47 bins of width 0.2 mag between @xmath49 and @xmath50 , spaced 0.1 mag apart ( adjacent bins overlap ) .
as described above , the xd fits for all but the brightest bin are initialized using the best - fit parameters for the previous bin .
each bin uses the full set of 103,601 redshift @xmath51 quasars . in each of 47 bins we fit 60 @xmath52-dimensional gaussians , yielding a total of @xmath53 - 1)$ ] parameters .
the @xmath0-only model has 59,173 parameters , the model that also uses the two uv bands has 101,473 parameters , the model that adds the four nir bands to the optical fluxes has 155,053 parameters , and the full uv-@xmath0-nir 11-dimensional model has 219,913 parameters . to obtain the total number of parameters for photometrically classifying quasars using _ xdqsoz _ , we need to add the number of parameters describing the stellar relative - flux density in 47 bins from the _ xdqso _ method to this number , amounting to 14,053 , 26,273 , 42,253 , and 61,993 parameters for the @xmath0 , @xmath0+uv , @xmath0+nir , and @xmath0+uv+nir models , respectively .
models including uv or nir data are trained using any available data , i.e. , any object with a measured flux in any of the bandpasses is included in the training set . in this section ,
we assess the performance of the xd technique for modeling the relative - flux redshift distribution of quasars , and provide examples which demonstrate that the xd technique produces excellent fits to the data .
we demonstrate that the xd method does an excellent job of empirically calibrating the color
redshift relation ; the ability of the xd technique to model the relative - flux density of quasars in the current _ xdqsoz _ context is excellent as well , but it is very similar to the performance in the _ xdqso _ context and we refer the reader to @xcite for a discussion of this performance . [
fig : allzexfit ] shows relative - flux redshift and color redshift diagrams of quasars for a single @xmath41-band magnitude bin .
the conditional distribution of relative - flux as a function of redshift is shown here ( although the model contains a full model of the density on this manifold ) ; this emphasizes what is new in _ xdqsoz _ as compared to _
we see that the xd technique is superb at capturing the complexity of the quasar color locus , even at higher redshifts where the data are sparse and noisy .
the locations where prominent emission lines cross the relevant _ sdss _ filters are indicated , and it is clear that this drives much of the structure in the color
redshift relation .
[ fig : allzexfitgalex ] shows similar relative - flux
redshift diagrams for the uv fluxes in the model containing both optical and uv data .
the agreement between the empirical model and the data is excellent .
these diagrams clearly demonstrate that the uv flux of @xmath54 and @xmath55 quasars , for fuv and nuv respectively , is suppressed because of absorption below the lyman limit ( @xmath56 ) by intervening systems @xcite .
uv observations are an excellent tool to distinguish @xmath57 quasars from @xmath58 quasars , which have degenerate @xmath0 colors and plague medium - redshift quasar selection ( e.g. , * ? ? ?
* ) , even at low uv signal - to - noise ratio .
[ fig : allzexfitukidss ] presents relative - flux redshift diagrams for the four nir fluxes in the _ xdqsoz _ model that contains optical and nir data .
the agreement between the _ xdqsoz _ model and the data is again excellent and the _ xdqsoz _ model captures all of the photometric redshift information contained in the nir .
we see that much of the variation in the color
redshift relation in the nir is driven by the h@xmath3 line ( see also * ? ? ?
* ; * ? ? ?
* ) . the model data comparisons given in this section are only a small fraction of the model - assessment diagnostics that we performed .
for example , we do not show the optical fits here in models that contain uv or nir data , nor do we show the uv models and nir models in the full @xmath0-uv+nir model , as all of these comparisons are very similar to the ones shown here .
we can use the _ xdqsoz _ model to photometrically classify and target quasars by calculating the probability that an object is a quasar based on its broadband fluxes .
the probability that an object is a quasar in a redshift range @xmath15 is obtained by integrating the probability that an object is a redshift @xmath16 quasar over redshift .
we start by using equation ( [ eq : quasardeltazprob ] ) @xmath59 where @xmath60 is the set of fluxes relative to @xmath61 and @xmath61 is the @xmath41-band flux of the object .
the normalization factor is given by @xmath62 because the apparent magnitude factor @xmath63 does not depend on redshift , the integral over redshift is only over the @xmath64 factor , which is modeled as a simple sum of gaussian distributions . for any given object we can simplify the mixture of @xmath52-dimensional gaussian distributions to a mixture of one - dimensional gaussian distributions for the redshift of the object .
first , we find the bin in the @xmath41-band magnitude that best matches the object s @xmath41-band magnitude and use the mixture - of - gaussians representation of the relative - flux redshift density in this bin .
assuming that the @xmath52-dimensional mixture of gaussians has amplitudes @xmath65 , means @xmath66 , and variance matrices @xmath67 , we can condition each of the components on the measured relative flux @xmath68 of the object and its uncertainty variance matrix @xmath69 to find ( e.g. , b of @xcite ) @xmath70 while the amplitudes of these one - dimensional gaussian distributions are given by the posterior probability that the object was drawn from component @xmath21 @xmath71 in these expressions @xmath72 and @xmath73 are the relative flux and the redshift parts of @xmath66 , respectively ; @xmath74 ; @xmath75 , @xmath76 , and @xmath77 are the relative - flux relative - flux , redshift relative - flux , and redshift redshift parts of @xmath67 , respectively ; and @xmath78 is the multivariate gaussian distribution .
@xmath79 includes the uncertainty variance matrix @xmath69 because the necessary uncertainty convolution simply reduces to adding the observational uncertainty variance matrix to the intrinsic variance matrix for each gaussian component . integrating this one - dimensional mixture of gaussian distributions over an arbitrary redshift range results in a sum over error functions . remembering that our model lives in @xmath80 redshift space @xmath81-\mathrm{erf}\left[\frac{\log z_{\mathrm{min } } - m_{z , k}}{\sqrt{2}\,\sigma_{z , k}}\right]\right)\ , , \end{split}\ ] ] where the error function erf@xmath82 \equiv 2 \int_0^x e^{-t^2}{\mathrm{d}}t\,/\sqrt{\pi}$ ]
. the first factor on the right - hand side of this equation is the integral over the entire redshift range @xmath83 $ ] , which simplifies to @xmath84 i.e. , the denominator in equation ( [ eq : alphazk ] ) .
we can compare the quasar probabilities obtained by integrating the _ xdqsoz _ model over redshift to those from the _ xdqso _ technique , which models the distribution of quasar fluxes in three wide redshift bins .
[ fig : xdqsovsxdqsoz ] shows the probabilities that 490,793 objects are medium - redshift ( @xmath85 ) quasars obtained by the two methods for objects in the _ sdss _ imaging stripe 82 .
it is clear that most of the objects cluster tightly around the one - to - one line and that the two models are essentially the same for this redshift range .
[ fig : xdqsovsxdqsozefficiency ] shows the efficiency of quasar targeting using both the _ xdqso _ and the _ xdqsoz _ method for targeting medium - redshift ( @xmath85 ) quasars .
this test uses a sample of medium - redshift quasars spectroscopically confirmed by _
boss_which also re - targets quasars previously identified in earlier surveys in stripe 82 .
this quasar sample is expected to be highly complete , because it was targeted using the superior imaging in stripe 82 where there is variability information @xcite and where a number of campaigns prior to _ boss _ have also obtained extensive spectroscopy .
the sample has on average of 30 @xmath86 quasars deg@xmath87 down to @xmath88 mag , which is close to the number expected from current quasar luminosity functions ( e.g. , * ? ? ?
we only use regions of stripe 82 that have more than 15 @xmath86 quasars deg@xmath87 . see @xcite and @xcite for a more detailed description of the _ boss _
quasar target selection in general and this test set in particular .
the top panel of [ fig : xdqsovsxdqsozefficiency ] shows selection based on _ sdss _
@xmath0 fluxes alone .
we see that the performance of the _ xdqsoz _ and the _ xdqso _
techniques is essentially identical .
the lower panels of this figure show how the selection improves when we add _ galex _ uv and _ ukidss _ _ las _ nir observations , both of which are available for essentially all objects in _ sdss _ stripe 82 .
the _ xdqso _
models with uv and nir data are models trained with these uv and nir fluxes in the broad redshift ranges used by _
xdqso_. for all intended purposes the _ xdqsoz _ technique performs identically to the _ xdqso _ technique for targeting medium - redshift quasars .
we have also checked the performance of the _ xdqsoz _ technique as compared to the kernel - density - estimation based photometric quasar classification technique of @xcite .
we find results that are similar to those for the _ xdqso _ technique as shown in 3 of @xcite : at low and medium redshift the _ xdqsoz _ technique performs slightly better than the _ xdqso _ technique ( and thus better than the kde technique ) , while at high - redshift ( @xmath89 ) _ xdqsoz _ performs slightly worse than _
xdqso_. this behavior is expected because the quasar training data do not include much data at high redshift .
we thus do not probe the color
redshift relation at high redshift as well as the kde approach , which included additional high - redshift data .
because we use the same stellar model as _ xdqso _ , the same problem with sampling regions of low stellar density that we encountered for _ xdqso _ persists for _
xdqsoz_. in summary , the _ xdqsoz _ technique performs almost identically to the _ xdqso _ method for photometrically classifying objects as quasars and thus for quasar targeting . _ xdqsoz _ has the advantage over _ xdqso _ and any other photometric quasar classification scheme that it can classify quasars in arbitrary redshift ranges `` on the fly '' ( i.e. , without retraining the model ) .
we have computed _ xdqsoz _ quasar probabilities for all 160,904,060 point sources with dereddened @xmath41-band magnitude between 17.75 and 22.45 mag in the 14,555 deg@xmath27 of imaging from sdss data release 8 @xcite in three redshift ranges ( 0.3 @xmath90 , @xmath91 , and @xmath37 ) .
[ fig : xdqsoz_imag ] shows the apparent @xmath41-band magnitude distribution of all of the objects with @xmath92 mag in the expected _ boss _ spectroscopic footprint @xcite with _
probability larger than 0.5 over the specified redshift range .
these apparent - magnitude distributions are smooth and well - behaved for low and medium redshifts .
they also agree at the bright end with number counts derived from spectroscopic observations @xcite . at the faint end ( @xmath93 )
the @xmath41-band number counts start to decline due to increasing photometric uncertainties and incompleteness of the _ sdss _ imaging near the faint limit of _
sdss_. the _ boss _ aims to detect the baryon acoustic feature ( baf ) in the ly@xmath3 forest of background redshift @xmath86 quasars .
not all quasars contribute equally to this measurement and , as shown by @xcite and @xcite , both brighter quasars and quasars near redshift @xmath94 are the most valuable . combining ly@xmath3 baf weights with the quasar probabilities as a function of redshift produced by _ xdqsoz _ , we can calculate the expected value of a quasar for the ly@xmath3 baf measurement . defining a value function @xmath95 , where @xmath96 is the dereddened @xmath96-band magnitude of the object , the expected value of an object is @xmath97 by targeting objects with the highest expected value for a particular ly@xmath3 baf survey which is dependent on the exact observational characteristics of that survey
we could optimize the targeting of quasars for that baf measurement .
the top panel of [ fig : xdqsozvsvalue ] shows the number of medium - redshift quasars found by applying this value - based targeting for _ boss _ using the value function of @xcite .
value - based targeting finds about 1 quasar deg@xmath87 less than targeting based on the ranked medium - quasar probability list .
the bottom panel shows that value - based targeting finds as much value as the straight probability - based targeting
but not more such that the baf measurement based on both samples should be equally precise .
straight probability - based targeting thus finds the same value while assembling a larger overall quasar sample .
in addition , value - based targeting optimizes one experiment in a specific survey , whereas straight probability - based targeting returns information that is broadly applicable to a range of experiments and a range of surveys .
thus , in general , there is little to be gained from pursuing value - based targeting for _
boss_. to investigate whether the _ xdqsoz _ quasar selection technique is limited by contamination from galaxies that appear point - like at the faint flux levels to which we push quasar classification , we look at the fraction of objects that appear point - like in a single _ sdss_imaging - pass but are extended in co - added data on _ sdss _ imaging stripe 82 .
we match the point sources in a typical _ sdss _ imaging run to the co - added galaxy catalog on stripe 82 @xcite and assess the fraction of point sources that are extended in the co - added data as a function of the @xmath41-band magnitude .
this is shown in the top panel of [ fig : pointgals ] .
we see that the fraction of point sources that are extended in the co - added imaging is only a few percent at relatively bright magnitudes , but almost reaches 50percent at @xmath98 mag .
to assess whether these point - like galaxies are a significant contaminant for the _ xdqsoz _ quasar selection , we calculate their quasar probabilities ( over all redshifts ) .
the fraction of point - like sources that are extended in the co - added imaging and that have quasar probabilities larger than 0.5 is shown as a function of the @xmath41-band magnitude in the lower panel of [ fig : pointgals ] . for comparison ,
the fraction of all point sources with quasar probability larger than 0.5 is shown as the dashed curve .
point - like galaxies make up only a small ( @xmath99percent ) fraction of _ xdqsoz_-selected photometric quasars .
however , because galaxies unlike stars cluster similarly to quasars , even this small contamination fraction might significantly degrade precision quasar - clustering measurements without improved star
galaxy separation or proper modeling . given the rising fraction of point - like galaxies with increasing magnitude in the top panel of [ fig : pointgals ] ,
point - like galaxies are likely to be the major contaminant for quasar selection at @xmath100 mag .
we can use the _ xdqsoz _ flux redshift density model to derive full posterior probability distributions for the redshift of photometric quasars taking the photometric uncertainties of the object fully into account . because the main advantages of the _ xdqsoz _ technique for photometric redshift estimation are that it a ) returns full pdfs and b )
allows auxiliary data such as that furnished by uv and nir surveys to be included , we focus on those points here .
@xmath0-only photometric quasar redshifts suffer from various degeneracies that are inherent to the @xmath0 filter system .
while including appropriate apparent - magnitude dependent redshift priors as we do here can partially relieve these degeneracies somewhat , no photometric redshift technique can entirely remove these degeneracies and _ xdqsoz _ is no exception ( as we will see below ) . crucially ,
even low signal - to - noise ratio uv and nir data _ can _ cleanly resolve these degeneracies . for each object the posterior probability distribution for its redshift , based on its measured broadband fluxes ,
is calculated by finding the apparent - magnitude bin that best matches the object s dereddened @xmath41-band magnitude .
the posterior probability distribution is given by the mixture of 60 one - dimensional gaussian distributions , with means , variances , and amplitudes given in equation ( [ eq : means ] ) , ( [ eq : vars ] ) , and ( [ eq : alphazk ] ) , respectively .
this posterior probability distribution can be calculated based on @xmath0 fluxes , or with additional uv or nir information if available .
we show four examples of such redshift pdfs in [ fig : predictdr7qso ] for objects from the _ sdss
_ dr7 quasar catalog that have measurements in all of the uv and nir filters .
these objects are from the sample used to train the flux
redshift quasar model , but , as we discuss in more detail below , we nevertheless believe that they provide an adequate representation of the performance of the _ xdqsoz _ technique .
these objects are chosen to demonstrate the power and weaknesses of the @xmath0 , uv , and nir data for photometric redshift estimation , and are therefore not a random subset of the data .
we discuss the overall performance below .
the top left panel shows an example where the @xmath0 fluxes suffice to accurately and precisely measure the redshift , and how the ( relatively high signal - to - noise ratio ) uv and nir measurements tighten the pdf significantly .
the top right panel shows an example where the extremely low uv flux basically vetoes the low - redshift peak that is present in the @xmath0-only redshift pdf . if one were to use a simple non - detection _ galex _
catalog at 5@xmath101 this result would not have been clear , because this object could have had the mean @xmath57 uv flux and still not be detected by _ galex_. the ability of the _ xdqsoz _ technique to use and interpret low signal - to - noise ratio data is therefore crucial in this example .
a weakness of the auxiliary uv data is apparent in the lower left panel .
here we see a @xmath102 quasar that is much brighter than the average quasar at this redshift in the uv , such that the addition of the uv data mistakenly chooses the low - redshift peak of the degenerate @xmath0 redshift pdf . however , the nir data are able to overcome this error and the addition of all the data confidently assigns this object a close - to - correct redshift . the lower right panel of [ fig : predictdr7qso ] shows another amusing example .
in addition to testing the _ xdqsoz _ technique using the _ sdss _ dr7 quasar sample , we have also drawn a sample of quasars located in the _ sdss _ imaging in stripe 82 discovered as part of the _ 2slaq _ survey @xcite and _ boss _ @xcite .
these quasars are generally fainter than the _ sdss _ quasars and therefore they represent a stringent , independent test of the _ xdqsoz _ technique s ability to return accurate redshift pdfs at faint magnitudes .
we have specifically selected all @xmath103 quasars with dereddened @xmath104 mag from the _ 2slaq _ sample and all quasars on the _ sdss _ imaging stripe 82 newly discovered by _
boss _ and use their single - pass _ sdss _ photometry .
because most of these objects lie in the _ sdss _ equatorial stripe , many of them have measurements from _ galex _ and _ ukidss _ _ las_. [ fig : predict2slaqboss ] shows posterior probability distributions for the redshift of two objects from the _
2slaq _ catalog and two from the _ boss _ sample .
the trends that were apparent for the _ sdss _
dr7 quasars in [ fig : predictdr7qso ] are also evident for these fainter objects .
the uv and nir fluxes for these objects have , in general , been measured much less precisely than those discussed above , but the auxiliary data still provide valuable extra information about the redshift . while most of the examples in s [ fig : predictdr7qso ] and [ fig : predict2slaqboss ] have a considerable posterior probability mass associated with the correct redshift , even when multiple peaks are present in the redshift pdf , this situation is generic in that inspections of many redshift pdfs show that it is rare to have no posterior probability mass associated with the spectroscopic redshift . as a simple statistic for the degeneracy in the redshift pdf we examine the number of distinct peaks as a function of redshift .
a single peak in the pdf is defined here as the widest contiguous region where the pdf is above the uniform distribution between redshift 0.3 and 5.5 ( i.e. , flat in redshift ) .
the top panel of [ fig : peaks ] shows the average number of such peaks as a function of redshift .
this statistic clearly shows the main degeneracies of @xmath0-based photometric quasar redshifts .
basically the entire @xmath105 region , a region around @xmath106 , and the @xmath107 redshift range are degenerate .
higher redshift quasars are readily identified as such using the @xmath0 colors ( e.g. , * ? ? ?
* ) . from the lower panels we see that the addition of uv and nir data softens all of these degeneracies .
essentially no degeneracies remain using the combination of all the uv , optical , and nir data ( lower panel of [ fig : peaks ] ) . additionally , requiring that distinct peaks in the photometric - redshift pdf need to have a minimum integrated probability ( e.g. , defining a peak as a contiguous region above the uniform distribution with @xmath108 integrated redshift probability ) gives similar results for the number of peaks and the improvement when adding uv and nir data .
[ fig : zspeczphotdr7qso ] shows the traditional spectroscopic - redshift vs. photometric - redshift diagram for quasars in the _ sdss _
dr7 quasar sample , for various combinations of wavelength regimes .
the right panels restrict the sample to those objects for which the redshift pdf has only a single peak and as such can be accurately described by a single photometric redshift ( plus uncertainty ) . in the top left panel all of the @xmath0-related degeneracies are clearly present and the right panel shows that by restricting the sample to single - peaked pdfs most of these degeneracies vanish , albeit at the cost of entire redshift ranges most notably redshift - range @xmath109 quasars .
the addition of uv and especially that of nir observations a ) greatly reduces the degeneracies as witnessed by the diminishing structure in the spectroscopic vs. photometric redshift plane and the increasing fraction of objects with a single peaked redshift pdf , and b ) significantly reduces the scatter . in the individual panels we report the number of 4@xmath101 outliers rather than the number of @xmath110 objects ; the latter number is somewhat meaningless without comparing it to the scatter , but to guide the eye we have included the @xmath111 lines .
the scatter is calculated without outlier - rejection .
we note that here and in the test below
the distribution of the @xmath41-band magnitude is unchanged when restricting the sample to objects with measured _ galex _ or _ ukidss _ fluxes ; restricting to objects with nir fluxes actually creates a fainter sample , because many faint quasars in the _ sdss _ imaging stripe 82 have been observed by _ ukidss _ _ las _ , while many brighter quasars are located outside of the _ ukidss _ _ las _ footprint . with the addition of uv and nir data most objects have accurate and precise single - peaked photometric redshifts over the entire @xmath112 redshift range : 97percent of all objects with uv and nir data and 99percent of the subset with single - peaked redshift pdfs have photometric redshifts within @xmath113 ; for @xmath114 these numbers are 84percent and 86percent respectively .
this is a significant improvement over @xmath0-only photometric redshifts , where we find 86percent of objects within @xmath113 .
similarly , @xcite found 83percent of objects in this range .
photometric and spectroscopic redshifts in [ fig : zspeczphotdr7qso ] are compared for objects in the sample used to train the _
xdqsoz _ technique .
as such , one might object that this is not a fair representation of the performance of the _ xdqsoz _ technique .
but , the relationship between the photometrically estimated redshift pdf and the spectroscopic redshift of a training object for _ xdqsoz _ is through the many - parameter flux redshift density model .
this model includes reweighting objects in the training set according to a redshift prior .
there is therefore no direct connection between output photometric redshifts and input spectroscopic redshifts as there is , for example , in nearest - neighbor approaches to photometric redshift estimation ( @xcite ) .
the fact that [ fig : zspeczphotdr7qso ] contains all of the expected redshift degeneracies for @xmath0-based photometric redshifts is further proof of this independence : if there were a dependent connection we would not suffer from these degeneracies . to further test this issue
we have divided our sample into a 90percent training sample and a 10percent test sample , as described above in section [ sec : modelconstruction ] .
we redo the spectroscopic - redshift vs. photometric - redshift comparison for the 10percent sample using the model trained in the 90percent of remaining data
the results are in [ fig : zspeczphottestqso ] . because the 10percent sample is much smaller than the full _ sdss _
dr7 quasar sample the statistics are noisier , but the trends are the same as in [ fig : zspeczphotdr7qso ] . to test the _ xdqsoz _
technique at fainter magnitudes , in [ fig : zspeczphot2slaqboss ] we compare spectroscopic redshifts to photometric redshifts for @xmath115 objects in the _ sdss _
dr7 quasar catalog and for objects in the combined _
2slaq _ and _ boss _ sample .
the trends in this figure are the same as those for the brighter _ sdss _ quasar sample and the scatter is somewhat larger ; however , the photometric redshifts remain clustered around the spectroscopic redshifts with no discernible bias . even for the faint _
2slaq _ and _ boss _ sample , the addition of low signal - to - noise ratio uv and nir data leads to a significant increase in accuracy and precision . using the technique described in this section we have computed photometric redshifts for all point sources in the expected _ boss_spectroscopic footprint with _ xdqsoz _
quasar probabilities larger than 0.5 and @xmath116 mag .
the distribution of peaks of the photometric - redshift distribution for these objects is shown in [ fig : xdqsoz_redshift ] in a few apparent - magnitude bins ( those bins from [ fig : redshiftprior ] that lie within the @xmath116 apparent - magnitude range .
the overall shape of the redshift distribution in each @xmath41-band bin is similar to the redshift prior calculated from the @xcite luminosity - function model .
however , the redshift - dependent efficiency of photometric quasar classification and redshift estimation is apparent in this comparison and the low classification efficiency at @xmath117 depresses the distribution in that range while increasing the significance of the @xmath118 peak .
all of the results in this section have assumed the @xcite redshift prior , shown in [ fig : redshiftprior ] .
using the difference between the hopkins , richards , & hernquist ( 2007 ) and @xcite redshift prior , given in the bottom panel of [ fig : redshiftprior ] , we can assess the difference in photometric redshift distribution when using these two alternatives to the quasar luminosity function .
the bottom panel of [ fig : redshiftprior ] shows that the only significant difference between these two models is at relatively high redshift ( @xmath119 ) and near the _ sdss _ detection limit ( @xmath93 mag ) .
strongly single - peaked photometric redshift distribution functions , such as many of those shown in s [ fig : predictdr7qso ] and [ fig : predict2slaqboss ] , are not affected by even order - of - magnitude changes in the redshift prior , especially when uv or nir data are available .
it is clear from s [ fig : zspeczphotdr7qso ] and [ fig : zspeczphot2slaqboss ] that , on average , the influence of a different redshift prior will be limited , as the main differences lie at higher redshift , where the _ sdss
_ colors provide relatively unambiguous photometric redshifts ( as shown by the lack of degeneracies at higher redshift in the photometric versus spectroscopic redshift plane ) .
as the photometric redshift distributions are only marginally affected by the use of a different prior , classification based on integration over these redshift pdfs also does not depend strongly on the details of the redshift prior .
we have previously discussed other photometric redshift estimation techniques for quasars in section [ sec : general ] . comparing [ fig : zspeczphotdr7qso ] to similar diagrams in @xcite we see , at least qualitatively , that the _ xdqsoz _ technique performs similarly when applied to the @xmath0 fluxes of bright , high signal - to - noise ratio objects .
we did not expect to perform better as the near - degeneracies in the @xmath0 color redshift plane are real and the quasar locus is broad .
the advantage of the _ xdqsoz _ technique over these other techniques is that it can be applied to faint objects and that it can incorporate uv and nir observations , even at low signal - to - noise ratio , to improve photometric redshift estimation and quasar classification .
no other method exists to calculate photometric quasar probabilities over arbitrary redshift ranges . by comparing with state - of - the - art photometric quasar classification using kernel - density estimation or gaussian mixture density deconvolution @xcite
, we have shown that the photometric quasar probabilities obtained by integrating the photometric redshift pdf over redshift are as good as those trained on the redshift range in question .
two additional sources of information relevant to photometric quasar classification and redshift estimation stand out as the next steps toward a full quasar model , although neither of these is currently available over the large areas of the sky surveyed by projects such as the _ sdss _ : photometric variability and differential - chromatic - refraction - induced astrometric offsets for quasars . of these
, photometric variability is the easiest to include in the quasar classification technique discussed here , as we can ignore the redshift information contained in the variability , because this information seems to be limited @xcite . as such , a photometric variability likelihood for quasars and stars
could be multiplied with the flux
redshift likelihood employed and modeled here to perform simultaneous color and variability selection .
the combination of photometric variability and color information will lead to accurate photometric quasar classification and redshift estimation in the _ lsst _ era .
the strong spectral features of quasars induce positional offsets to standard differential - chromatic - refraction corrections @xcite .
these positional offsets are redshift dependent much as quasar colors are redshift dependent because of spectral features moving through individual filters ( see , e.g. , [ fig : allzexfit ] ) .
thus , these offsets could be used to break redshift degeneracies . accomplishing this in the _ xdqsoz _ flux
redshift density context necessitates adding the astrometric offsets into the density model . because the astrometric offsets are zenith angle dependent
, these models would have to be constructed for a range of airmasses , or airmass could be added as an additional dimension .
as astrometric redshifts are a subtle and difficult - to - measure effect , the deconvolution aspect of the xd density - estimation technique could be useful .
the technique described in this article is a step toward a generalized method for object classification and characterization from broadband photometric data , which will become increasingly relevant in this era of major wide - field imaging surveys .
while our quasar model includes redshift in addition to the broadband fluxes of an object , our star model does not because stars do not possess a cosmological redshift .
stars are characterized by other properties e.g . ,
distance and metallicity that are often estimated photometrically ( e.g. , * ? ? ?
* ; * ? ? ?
as we are interested here in quasar classification and characterization , our model implicitly marginalized over stellar properties .
however , as part of a general object classification pipeline , these properties should be included and the technique developed in this paper could be applied .
more importantly , the general framework outlined in section [ sec : general ] and the specific implementation in sections [ sec : xdqsoz ] and [ sec : targeting ] show that we can perform classification when different models are characterized by different parameters even different _ numbers _ of parameters .
this aspect is especially relevant in the context of quasar selection based on variability .
quasar variability is commonly modeled as a stochastic gaussian process @xcite characterized by a small number of parameters .
recently it has been shown that this framework allows for a clean selection of quasars because most stars the main contaminants for
quasar targeting currently in general do not vary over long time baselines . in this type of selection , however , quasars and stars are often modeled ( or fit ) using the same stochastic model , which is inappropriate for the stars ( @xcite ; however , see @xcite ) .
the use of a stochastic model for variability - based star
quasar separation is particularly problematic for rr lyrae stars a common contaminant in color - based classification of quasars in some redshift ranges .
rr lyraes are known to vary periodically rather than stochastically . in the framework we use in this article all classes of objects
can be described using models appropriate for the class e.g .
, stochastically varying objects with a cosmological redshift for the quasars and non - variable sources for most stars
because object classification only uses _ marginalized _ probabilities , that is , probabilities marginalized over the internal properties of each class ( cf . equation [ [ eq : quasardeltazprob ] ] ) . describing each class with a model appropriate for
that class should lead to better classification and simultaneous object classification of sources into _ all _ classes .
as photometric quasar classification moves to ever fainter flux levels , contamination from point - like galaxies becomes increasingly important . as discussed by @xcite ,
unresolved galaxies are implicitly taken into account in our model because our training set of `` stars '' is actually a set of non - variable point - like objects that therefore includes faint galaxies .
this model of galaxies again implicitly marginalizes over galaxy properties , most notably the redshift of the galaxy .
photometric redshift estimation for galaxies is closely related to obtaining photometric redshifts for quasars , but the galaxy photometric redshift techniques tend to rely more on templates in their model building , while quasars are modeled in a more empirical manner ( e.g. , * ? ? ? * ) . however , this distinction is not fundamental and the general framework discussed here still applies .
template - based models are just another way of obtaining the probability @xmath120 or @xmath121 .
the _ xdqsoz _ model of section [ sec : xdqsoz ] also contains the distribution of broadband fluxes as a function of redshift @xmath122 such as is used to compute mean quasar color tracks .
this probability density is obtained from the full _ xdqsoz _ model flux
redshift density by conditioning on redshift . for the relative flux
this leads to a mixture of gaussians with means , variances , and amplitudes given by expressions similar to those in equations ( [ eq : means ] ) and ( [ eq : alphazk])essentially , relative flux and redshift need to be interchanged in those equations
. properties of this distribution can be calculated from the mixture of gaussians .
for example , the mean quasar relative flux ( or color ) as a function of redshift is obtained by weighting the means of the gaussian components using the redshift - dependent amplitudes .
code to calculate the mean quasar color track is included in the package described in the .
in this article we have introduced a new approach to photometric quasar classification that can simultaneously classify quasars and characterize their redshifts based on broadband photometry .
this technique , _ xdqsoz _ , is an extension of the _ xdqso _ technique of @xcite that adds the unknown redshift as an extra parameter to the quasar model to obtain the likelihood @xmath123 that is central to both quasar classification and photometric redshift estimation .
we have shown that this combined approach is both the best current quasar classification technique
it has similar performance as the _ xdqso _ method and a competitive photometric redshift method .
compared to other approaches to photometric redshift estimation for quasars it has the advantage that it can incorporate additional uv and nir data , even at low signal - to - noise ratio , and can be extended to fainter flux levels where photometric uncertainties are significant . using samples of quasars drawn from the _ sdss
_ , _ 2slaq _ , and _ boss _ spectroscopic catalogs we have demonstrated this increased performance down to @xmath124 mag .
the addition of uv and nir data to the photometric redshift estimation problem essentially breaks all of the redshift degeneracies inherent to the @xmath0 filter set .
code to use the _ xdqsoz _ technique for classification and redshift estimation , including the ability to calculate full posterior probability distributions for the redshift , are made publicly available .
this code is briefly described in the .
it is a pleasure to thank the anonymous referee and paul martini , gordon richards , and david weinberg for helpful comments and discussions
. j.b . and d.w.h .
were partially supported by nasa ( grant nnx08aj48 g ) and the nsf ( grant ast-0908357 ) .
a.d.m . acknowledges support under the nasa adap program ( grant nnx08aj28 g ) .
j.f.h acknowledges support provided by the alexander von humboldt foundation in the framework of the sofja kovalevskaja award endowed by the german federal ministry of education and research .
d.w.h . and a.d.m .
are research fellows of the alexander von humboldt foundation of germany .
we gratefully acknowledge nasa s support for construction , operation , and science analysis for the _ galex _ mission , developed in cooperation with the centre national detudes spatiale of france and the korean ministry of science and technology . this work is based in part on data obtained as part of the ukirt infrared deep sky survey ( _ ukidss _ ) .
funding for the sdss and sdss - ii has been provided by the alfred p. sloan foundation , the participating institutions , the national science foundation , the u.s .
department of energy , the national aeronautics and space administration , the japanese monbukagakusho , the max planck society , and the higher education funding council for england .
the sdss web site is http://www.sdss.org/. the sdss is managed by the astrophysical research consortium for the participating institutions .
the participating institutions are the american museum of natural history , astrophysical institute potsdam , university of basel , university of cambridge , case western reserve university , university of chicago , drexel university , fermilab , the institute for advanced study , the japan participation group , johns hopkins university , the joint institute for nuclear astrophysics , the kavli institute for particle astrophysics and cosmology , the korean scientist group , the chinese academy of sciences ( lamost ) , los alamos national laboratory , the max - planck - institute for astronomy ( mpia ) , the max - planck - institute for astrophysics ( mpa ) , new mexico state university , ohio state university , university of pittsburgh , university of portsmouth , princeton university , the united states naval observatory , and the university of washington .
sdss - iii is managed by the astrophysical research consortium for the participating institutions of the sdss - iii collaboration including the university of arizona , the brazilian participation group , brookhaven national laboratory , university of cambridge , university of florida , the french participation group , the german participation group , the instituto de astrofisica de canarias , the michigan state / notre dame / jina participation group , johns hopkins university , lawrence berkeley national laboratory , max planck institute for astrophysics , new mexico state university , new york university , the ohio state university , the penn state university , university of portsmouth , princeton university , university of tokyo , the university of utah , vanderbilt university , university of virginia , university of washington , and yale university .
the _ xdqsoz _ code for target selection , classification , and photometric redshift estimation is publicly available at _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ http://www.sdss3.org/svn/repo/xdqso/tags/v0_6/doc/build/html/index.html .
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ the _ xdqso_/_xdqsoz _
package contains routines for quasar classification using the _ xdqso _ and _ xdqsoz _ techniques .
it also contains code to calculate posterior probability distributions for the quasar redshift of objects based on input ` psfflux ` and ` psfflux_ivar ` ; these can be found in standard _ sdss _
data files such as the ` sweeps ' files the _ xdqsoz _ models for the quasar color redshift density are contained in the ` data/ ` directory .
they are in the form of fits files containing the xd models for all of the bins in apparent magnitude , with one file for each combination of _ sdss _ with _ galex _ and _ ukidss_. each fits file contains 47 extensions , where extension @xmath21 contains a structure with the amplitudes ( tag `` ) , means ( tag `` ) , and covariance matrices ( tag `` ) for bin @xmath21 in @xmath41-band magnitude .
the zeroth dimension of the gaussian represents the natural logarithm of the redshift , followed by _
sdss _ , _ galex _ , and _ ukidss _ fluxes ( where relevant ) in this order , and ordered as nuv / fuv for the _ galex _ , and yjhk for the _ ukidss_. abazajian , k. n. , et al . , 2009 , , 182 , 543 abell , p. a. , et al . , 2009 , the lsst science book , arxiv:0912.0201v1 [ astro - ph ] aihara , h. , et al . , 2011 , , 193 , 29 assef , r. j. , et al . , 2010 , , 713 , 970 atlee , d. w. & gould , a. , 2007 , , 664 , 53 baldwin , j. a. , 1977 , , 214 , 679 ball , n. m. , brunner , r. j. , myers , a. d. , strand , n. e. , alberts , s. l. , tcheng , d. , & llor , x. , 2007 , , 663 , 774 ball , n. m. , brunner , r. j. , myers , a. d. , strand , n. e. , alberts , s. l. , & tcheng , d. , 2008 , , 683 , 12 baum , w. a. , 1962 , in problems of extra - galactic research , proceedings from iau symposium no . 15 , edited by mcvittie , g. c. , 390 bentez , n. , 2000 , , 536 , 571 bianchi , l. , efremova , b. , herald , j. , girardi , l. , zabot , a. , marigo , p. , & martin , c. , 2011 , , 411 , 2770 , j. , hogg , d. w. , & roweis , s. t. 2009 , ann .
stat . 5 , 2b , 1657 , arxiv:0905.2979 bovy , j. , et al . , 2011 , , 729 , 141 bowen , d. v. , et al . , 2006 , , 645 , l105 budavri , t. , et al . , 2001 , , 122 , 1163 butler , n. r. & bloom , j. s. , 2010 , , 141 , 93 casali , m. , et al . , 2007 , , 467 , 777 chiu , k. , richards , g. t. , hewett , p. c. , & maddox , n. , 2007 , , 375 , 1180 connolly , a. j. , csabai , i. , szalay , a. s. , koo , d. c. , kron , r. g. , & munn , j. a. , 1995 , , 110 , 2655 croom , s. m. , warren , s. j. , & glazebrook , k. , 2001 , , 328 , 150 croom , s. m. , et al . , 2009 , , 392 , 19 dabrusco , r. , longo , g. , & walton , n. a. , 2009 , , 396 , 223 dye , s. , et al
. , 2006 , , 372 , 1227 eisenstein , d. , et al . , 2011 , , 142 , 72 fan , x. , et al . , 1999 , , 118 , 1 francis , p. j. , nelson , b. o. , & cutri , r. m. , 2004 , , 127 , 646 fukugita , m. , 1996 , , 111 , 1748 giannantonio , t. , scranton , r. , crittenden , r. g. , nichol , r. c. , boughn , s. p. , myers , a. d. , & richards , g. t. , 2008 , , 77 , 123520 giannantonio , t. , et al .
, 2006 , , 74 , 063520 glikman , e. , helfand , d. j. , & white , r. l. , 2006 , , 640 , 579 gunn , j. e. , et al . , 1998 , , 116 , 3040 gunn , j. e. , et al . , 2006 , , 131 , 2332 hambly , n. c. , et al . , 2008 , , 384 , 637 hennawi , j. f. , et al . , 2006a , , 131 , 1 hennawi , j. f. , et al
. , 2006b , , 651 , 61 hennawi , j. f. & prochaska , j. x. , 2007 , , 655 , 735 hennawi , j. f. , et al . , 2010 , , 719 , 1672 hewett , p. c. , warren , s. j. , leggett , s. k. , & hodgkin , s .t . , 2006 , , 367 , 454 hodgkin , s. t. , irwin , m. j. , hewett , p. c. , & warren , s. j. , 2009 , , 394 , 675 hogg , d. w. , finkbeiner , d. p. , schlegel , d. j. , & gunn , j. e. , 2001 , , 122 , 2129 hook , j. m. , mcmahon , r. g. , boyle , b. j. , & irwin , m. j. , 1994 , , 268 , 305 hopkins , p. f. , richards , g. t. , & hernquist , l. , 2007 , , 654 , 731 hutchings , j. b. & bianchi , l. , 2010 , , 140 , 1987 ivezi , . , et al . , 2003 , mem
. soc . astron .
italiana , 74 , 978 ivezi , .
, et al . , 2004 ,
an , 325 , 583 ivezi , . , et al . , 2007 , , 134 , 973 ivezi , . , et al . , 2008 , 684 , 287 jimenez , r. , spergel , d. n. , niemack , m. d. , menanteau , f. , hughes , j. p. , verde , l. , & kosowsky , a. , 2009 , , 181 , 439 juri , m. , et al .
, 2008 , , 673 , 864 kaczmarczik , m. c. , richards , g. t. , mehta , s. s. , & schlegel , d. j. , 2009 , , 138 , 19 kelly , b. c. , bechtold , j. , & siemiginowska , a. , 2009 , , 698 , 895 kozowski , s. , et al . , 2010a , , 708 , 927 kozowski , s. , et al . , 2010b , , 716 , 530 lawrence , a. , et al . , 2007 , , 379 , 1599 lupton , r. , et al . , 2001 ,
aspc , 238 , 269 martin , d. c. , et al .
, 2005 , , 619 , l1 mcdonald , p. & eisenstein , d. j. , 2007 , , 76 , 063009 macleod , c. l. , et al .
, 2011a , , 721 , 1014 macleod , c. l. , et al . , 2011b , , 728 , 26 mcquinn , m. & white , m. , 2011 , , 415 , 2257 maddox , n. & hewett , p. c. , 2006 , , 367 , 717 mller , p. & jakobsen , p. , 1990
, , 228 , 299 morrissey , p. , et al .
, 2007 , , 173 , 682 mortlock , d. j. , et al . , 2012 , , 419 , 390 myers , a. d. , brunner , r. j. , richards , g. t. , nichol , r. c. , schneider , d. p. , vanden berk , d. e. , scranton , r. , gray , a. g. , & brinkmann , jon , 2006 , , 638 , 622 myers , a. d. , brunner , r. j. , nichol , r. c. , richards , g. t. , schneider , d. p. & bahcall , n. a. , 2007a , , 658 , 85 myers , a. d. , brunner , r. j. , nichol , r. c. , richards , g. t. , schneider , d. p. & bahcall , n. a. , 2007b , , 658 , 99 myers , a. d. , richards , g. t. , brunner , r. j. , schneider , d. p. , strand , n. e. , hall , p. b. , blomquist , j. a. , & york , d. g. , 2008 , , 678 , 635 myers , a. d , white , m. , & ball , n. m. , 2009 , , 399 , 2279 oke , j. b. & gunn , j. e. , 1983 , , 266 , 713 padmanabhan , n. , et al . , 2008 , , 674 , 1217 palanque - delabrouille , et al . , 2011 , , 530 , a122 peth , m. a. , ross , n. p. , & schneider , d. p. , 2011
, , 141 , 105 picard , a. & jakobsen , p. , 1993
, , 276 , 331 pier , j. r. , et al . , 2003 , , 125 , 1559 pogge , r. w. , et al . , 2010 , in ground - based and airborne instrumentation for astronomy iii . edited by mclean , i. s. , ramsay , s. k. , & takami , h. , 7735 , 77350a prochaska , j. x. & hennawi , j. f. , 2009 , , 690 , 1558 richards , g. t. , et al .
, 2001a , , 121 , 2308 richards , g. t. , et al .
, 2001b , , 122 , 1151 richards , g. t. , et al . , 2002 , , 123 , 2945 richards , g. t. , et al . , 2004 , , 155 , 257 richards , g. t. , et al . , 2006 , , 131 , 2766 richards , g. t. , et al .
, 2009a , , 180 , 67 richards , g. t. , et al .
, 2009b , , 137 , 3884 ross , n. p. , et al . , 2011 , , in press , arxiv:1105.0606 schlegel , d. j. , finkbeiner , d. p. , & davis , m. , 1998 , , 500 , 525 schmidt , k. b. , et al . , 2010 , , 714 , 1194 schneider , d. p. , et al . , 2010 , , 139 , 2360 scranton , r. , et al . ,
2005a , , 633 , 589 scranton , r. , et al .
, 2005b , arxiv : astro - ph/0508564 shen , y. , et al . , 2010 , , 719 , 1693 smith , j. a. , et al . , 2002 , , 123 , 2121 stoughton , c. , et al . , 2002 , , 123 , 485 suchkov , a. a. , hanisch , r. j. , & margon , b. , 2005 , , 130 , 2439 trammell , g. b. , vanden berk , d. e. , schneider , d. p. , richards , g. t .. hall , p. b. , anderson , s. f. , & brinkmann , j. , 2007 , , 133 , 1780 vanden berk , d. e. , et al . , 2004 , , 601 , 692 warren , s. j. , hewett , p. c. , & foltz , c. b. , 2000 , , 312 , 827 weinstein , m. a. , et al . , 2004 , , 155 , 243 welsh , b. y. , wheatley , j. m. , neil , j. d. , 2011 , , 527 , 15 wolf , c. , et al . , 2004 , , 421 , 913 worseck , g. & prochaska , j. x. , 2011 , , 728 , 23 wu , x .- b . & jia , z. , 2010 , , 406 , 1583 yip , c. w. , et al . , 2004 , , 128 , 2603 york , d. g. , et al . , 2000 , , 120 , 1579 62,628 objects ; _ top panel _ ) and _ ukidss _ _ las _ ( @xmath25 25,510 objects ; _ bottom panel _ ) .
see sections [ sec : galexdata ] and [ sec : ukidssdata ] for the number of objects in each individual bandpass .
the five - sigma detection limit is indicated.,scaledwidth=60.0% ] ) quasar probabilities computed using _
xdqso _ , that is , based on flux - density models in broad redshift ranges , and _
xdqsoz _ , i.e. , obtained by integrating flux
density models over the relevant redshift range , for 490,793 objects in _ sdss _ stripe 82 based on single - imaging - run flux measurements .
conditional 25 , 50 , and 75percent quantiles are shown . ] ) quasar selection efficiency for _ xdqso _ and _ xdqsoz _ as a function of target density for objects in _ sdss _ stripe 82 based on single - imaging - run flux measurements .
the top panel bases selection solely on _ sdss _
@xmath0 fluxes , the lower panels add _ galex _ nuv and fuv medium - deep measurements as well as _ ukidss _
yjhk photometry , both of which are available for almost all stripe-82 sources , through force - photometering _ galex _ and _ ukidss _ _ las _ imaging data at _ sdss _ positions . the 50percent selection efficiency is indicated and the @xmath0-only curve for _ xdqsoz _ is repeated in gray in each panel . ]
-band magnitude distribution of all point sources in the expected _ sdss - iii _ _ boss _ footprint with _ xdqsoz _
quasar probability larger than 0.5 over the indicated redshift range and dereddened @xmath41 between 17.8 mag and 21.5 mag .
diamonds indicate number counts from the _ sdss _ spectroscopic survey @xcite.,scaledwidth=70.0% ] forest , while probability - based selection ranks on @xmath125 quasar ) .
the top panel shows the number of mid - redshift quasars found by each method as a function of the target density ; the bottom panel shows the value of the selected quasars .
note that some @xmath126 quasars are valuable for the lyman-@xmath3 forest baf measurement . ]
-band magnitude .
the bottom panel shows the fraction of such point - like galaxies that have an _ xdqsoz _
quasar probability ( over all redshifts ) larger than 0.5 .
the dashed curve in the bottom panel shows the fraction of all point - sources that have an _ xdqsoz _
quasar probability larger than 0.5 .
even though point - like galaxies start to dominate the number counts around @xmath41 = 22 mag , they only make up a small fraction of photometrically selected quasars at all magnitudes . ] .
the overall shape of the redshift distribution is similar to the prior distribution , except for the drop in @xmath127 due to the decreased efficiency of photometric quasar classification based on _ sdss _ photometry.,scaledwidth=70.0% ] | we describe a technique for simultaneously classifying and estimating the redshift of quasars .
it can separate quasars from stars in arbitrary redshift ranges , estimate full posterior distribution functions for the redshift , and naturally incorporate flux uncertainties , missing data , and multi - wavelength photometry .
we build models of quasars in flux redshift space by applying the _ extreme deconvolution _ technique to estimate the underlying density . by integrating this density over redshift
one can obtain quasar flux
densities in different redshift ranges .
this approach allows for efficient , consistent , and fast classification and photometric redshift estimation .
this is achieved by combining the speed obtained by choosing simple analytical forms as the basis of our density model with the flexibility of non - parametric models through the use of many simple components with many parameters .
we show that this technique is competitive with the best photometric quasar classification techniques which are limited to fixed , broad redshift ranges and high signal - to - noise ratio data and with the best photometric redshift techniques when applied to broadband optical data .
we demonstrate that the inclusion of uv and nir data significantly improves photometric quasar
star separation and essentially resolves all of the redshift degeneracies for quasars inherent to the @xmath0 filter system , even when included data have a low signal - to - noise ratio . for quasars
spectroscopically confirmed by the _
sdss _ 84 and 97percent of the objects with _
_ uv and _ ukidss _
nir data have photometric redshifts within 0.1 and 0.3 , respectively , of the spectroscopic redshift ; this amounts to about a factor of three improvement over @xmath0-only photometric redshifts .
our code to calculate quasar probabilities and redshift probability distributions is publicly available . |
since last many years , investigation about the nuclear equation of state ( neos ) at the extreme conditions of density and temperature has been one of the primary driving forces in heavy ion studies at intermediate energies .
the interest in low energies , however , are for isospin effects in fusion process @xcite .
these investigations has been performed with the help of rare phenomena such as multifragmentation , collective flow , particle production as well as nuclear stopping @xcite .
the relation between the nuclear eos and flow phenomena has been explored extensively in the simulations .
recently the analysis of transverse - momentum dependence of elliptical flow has also been put forwarded @xcite .
the elliptical flow is shaped by the interplay between the geometry and mean field and , when gated by the transverse momentum , reveals the momentum dependence of the mean field at supra - normal densities .
the parameter of the elliptic flow is quantified by the second - order fourier coefficient @xcite @xmath0 from the azimuthal distribution of detected particles at mid rapidity as @xmath1 where @xmath2 and @xmath3 are the @xmath4 and @xmath5 components of momentum .
the @xmath2 is in the reaction plane , while , @xmath3 is perpendicular to the reaction plane and @xmath6 is the azimuthal angle of emitted particles momentum relative to the x - axis .
the positive values of @xmath7 reflect preferential in - plane emission , while negative values reflect preferential out - of - plane emission .
the pulsating of sign observed recently at intermediate energies has received particular attention as it reflects the increased pressure buildup in the non isotropic collision zone @xcite .
+ after the pioneering measurements at saturne @xcite and bevalac @xcite , a wealth of experimental results have been obtained at bevalac and sis @xcite as well as at ags @xcite , sps @xcite and rhic @xcite . in recent years ,
the fopi , indra , and plastic ball collaborations @xcite are actively involved in measuring the excitation function of elliptical flow from fermi energies to relativistic one . in most of these studies ,
collisions of @xmath8 is undertaken .
interestingly , elliptical flow was reported to change from positive ( in - plane ) taken .
negative ( out - of - plane ) values around 100 mev / nucleon and maximum squeeze - out is observed around 400 mev / nucleon .
these observations were reported recently by us and others theoretically @xcite .
a careful analysis reveals that elliptic flow is very sensitive towards the choice of rapidity cut which makes the difference between spectator and participant matter . + the elliptical flow pattern of participant matter is affected by the presence of cold spectators @xcite .
when nucleons are decelerated in the participant region , the longitudinal kinetic energy associated with the initial colliding nuclei is converted into the thermal and potential compression energy . in a subsequent rapid expansion(or explosion ) , the collective transverse energy develops @xcite and many particles from the participant region get emitted into the transverse directions .
the particles emitted towards the reaction plane can encounter the cold spectator pieces and , hence , get redirected .
in contrast , the particles emitted essentially perpendicular to the reaction plane are largely unhindered by the spectators .
thus , for the beam energies leading to rapid expansion in the vicinity of the spectators , elliptic flow directed out of the reaction plane ( squeeze - out ) is expected .
this squeeze - out is related with the pace at which expansion develops , and is , therefore , related to the eos .
this contribution of the participant and spectator matter @xcite in the intermediate energy heavy - ion collisions motivated us to perform a detailed analysis of the excitation function of elliptical flow over different regions of participant and spectator matter . if excitation function is found to be affected by the different contributions it will definitely , also affect the in - plane to out - of - plane emission i.e. transition energy . + the rapidity distribution is an important parameter to study the participant - spectator contribution in the intermediate energy heavy - ion collisions @xcite . in this paper
, we will study the effect of participant and spectator matters in term of different rapidity distributions on the excitation function of elliptical flow .
attempts shall also be made to parameterize the transition energy in term of different rapidity bins .
+ the entire work is carried out in the framework of isospin - dependent quantum molecular dynamics ( iqmd ) @xcite .
the iqmd model is discussed in detail in the following section .
the iqmd model @xcite , which is an improved version of qmd model @xcite developed by j. aichelin and coworkers , then has been used successfully to various phenomena such as collective flow , disappearance of flow , fragmentation & elliptical flow @xcite .
the isospin degree of freedom enters into the calculations via symmetry potential , cross - sections and coulomb interaction @xcite .
the details about the elastic and inelastic cross - sections for proton - proton and neutron - neutron collisions can be found in ref .
+ in iqmd model , the nucleons of target and projectile interact via two and three - body skyrme forces , yukawa potential and coulomb interactions .
in addition to the use of explicit charge states of all baryons and mesons , a symmetry potential between protons and neutrons corresponding to the bethe- weizsacker mass formula has been included .
+ the hadrons propagate using classical hamilton equations of motion : @xmath9 with @xmath10 the baryon - baryon potential @xmath11 , in the above relation , reads as : @xmath12 here @xmath13 and @xmath14 denote the charges of @xmath15 and @xmath16 baryon , and @xmath17 , @xmath18 are their respective @xmath19 components ( i.e. 1/2 for protons and -1/2 for neutrons ) .
the parameters @xmath20 and @xmath21 are adjusted to the real part of the nucleonic optical potential . for the density dependence of nucleon optical potential ,
standard skyrme - type parameterization is employed .
the potential part resulting from the convolution of the distribution function with the skyrme interactions @xmath22 reads as : @xmath23 the two of the three parameters of equation of state are determined by demanding that at normal nuclear matter density , the binding energy should be equal to 16 mev .
the third parameter @xmath24 is usually treated as a free parameter .
its value is given in term of the compressibility : @xmath25 the different values of compressibility give rise to soft and hard equations of state .
it is worth mentioning that skyrme forces are very successful in analysis of low energy phenomena like fusion , fission and cluster - radioactivity . +
as noted @xcite , elliptical flow is weakly affected by the choice of equation of state . on the other hand , in the refs .
@xcite , the hard equation of state is used to study the directed as well as elliptical flow . for the present analysis ,
a hard ( h ) equation of state has been employed along with standard energy dependent cross - section . +
for the present analysis , simulations are carried out for thousand of events for the reaction of @xmath8 at semi - central geometry using a hard equation of state .
the whole of the analysis is performed for light charged particles ( lcps)[1@xmath26 a @xmath26 4 ] .
the reaction conditions and fragments are chosen on the basis of availability of experimental data @xcite . as noted in ref .
@xcite , the relativistic effects do not play role at these incident energies and the intensity of sub - threshold particle production is very small .
the phase space generated by the iqmd model has been analyzed using the minimum spanning tree ( mst ) @xcite method .
the mst method binds two nucleons in a fragment if their distance is less than 4 fm .
this is the one of the widely used method in the intermediate energy heavy - ion collisions .
however , some improvements like momentum and binding energy are also discussed in the literature @xcite . in recent years ,
more sophisticated and complicated algorithms are also available in the literature @xcite .
the entire calculations are performed at @xmath27 fm / c .
this time is chosen by keeping in view the saturation of the collective flow @xcite . + as our purpose of present study is to understand the effect of participant - spectator matter on the excitation function of elliptical flow , one will concentrate on the rapidity distribution only @xcite .
the important concept of spectators and participants in collisions was first introduced by bowman et al .
@xcite and later employed for the description of a wide - angle energetic particle emission by westfall et al .
the two nuclei slamming against one other can be viewed as producing cylindrical cuts through each other .
the swept - out nucleons or participants ( from projectile and target ) undergo a violent collision process . in the meantime , the remnants of the projectile and target continue with largely undisturbed velocities , and are much less affected by the collision process than the participant nucleons . on one hand , this picture is supported by features of the data @xcite and , on the other , by dynamic simulations @xcite . during the violent stage of a reaction , the spectators can influence the behavior of participant matter . following the same
, we are dividing the rapidity distribution in the different cuts in term of @xmath28 parameter , which is given as : @xmath29 where e(i ) and @xmath30 are respectively , the total energy and longitudinal momentum of @xmath15 particle . +
as we are interested in studying the effect of rapidity distribution on the incident energy dependence of elliptical flow , one has to understand the @xmath31 as a function of @xmath32 at different energies . for this , in fig.[fig:1 ]
, we display the rapidity distribution for the light charged fragments ( lcp s ) at different incident energies .
the rapidity distribution is found to vary drastically throughout the range from @xmath33 .
this is indicating the compressed or participant zone around @xmath34 value and decay into spectator zone towards both side of the @xmath34 value .
however around @xmath34 value , the region between @xmath35 to @xmath36 is specified as mid - rapidity region in the literature @xcite . the decay from @xmath35 towards negative side approach to target like spectator , while , other side is known as projectile like spectator .
interestingly , these regions are found to affect drastically with incident energy . with increase in incident energy , number of particles are increasing in the region from @xmath37 to @xmath38 , while decreasing away from this region on both side .
the inset in the figure is showing the clear view of change in behavior around @xmath37 , which is also true for other side also .
the rapidity distribution is also used as thermalization source in the literature many times @xcite . on the basis of this
, we have divided the rapidity distribution in five different zones out of which three are with including @xmath39(participant zone ) , while other two are with excluding this region ( target and projectile spectator).they are as : ( i ) from -0.1 @xmath26 @xmath40 @xmath26 0.1 with increment of 0.2 on both side upto -1.5
@xmath26 @xmath40 @xmath26 1.5 .
( ii ) from 0 @xmath26 @xmath40 @xmath26 0.1 with increment of 0.2 on latter side upto 0 @xmath26 @xmath40 @xmath26 1.5 . (
iii ) from -0.1 @xmath26 @xmath40 @xmath26 0 with decrement of 0.2 the on former side upto -1.5 @xmath26 @xmath40 @xmath26 0 .
( iv ) from @xmath40 @xmath41 0.1 with increment of 0.2 upto @xmath40 @xmath41 1.5 . and ( v)from @xmath40 @xmath42 -0.1 with decrement of 0.2 upto @xmath40 @xmath42 -1.5 .
+ the rapidity distribution dn / dy as a function of @xmath28 at different incident energies ranging from 50 to 1000 mev / nucleon for light charged particles ( lcps)[1@xmath26 a @xmath26 4 ] .
the inset is showing only one side of the rapidity distribution . ] the incident energy dependence of elliptical flow for lcp s collectively for projectile as well as target matter including mid - rapidity region .
the different lines are at different size of the rapidity bin , which includes the participant as well as spectator matter . ]
let us now understand the effect of these rapidity cuts on the excitation function of elliptical flow . in fig.[fig:2
] , we present the excitation function of elliptical flow for the first condition . this condition is the mixture of the participant as well as spectator zone from projectile as well as target matter .
there are two points to discuss here .
one is the change in the elliptical flow with incident energy and other is to see the effect of bin size on elliptical flow .
the elliptical flow evolves from a positive value , rotational - like , emission to an negative value , collective expansion , with increase in the incident energy . in other words ,
transition from the in - plane to out - of - plane takes place .
the energy at which this transition takes place is known as transition energy .
this transition is due to the competition between the mean field at low incident energy and nn collisions at high incident energies .
the incident energy dependence of directed flow also show such type of transitions from negative to positive value , which is know as balance energy @xcite .
after this transition , the strength of collective expansion overcomes the rotational like motion @xcite .
this leads to increase of out - of - plane emission towards a maximum around 400 mev / nucleon .
this maxima is further supported by the nuclear stopping at 400 mev / nucleon @xcite . beyond this energy , elliptical flow decreases indicating a transition to in - plane preferentially emission @xcite .
this rise and fall behavior of the elliptical flow in the expansion region is due to the variation in the passing time @xmath43 of the spectator and expansion time of the participant zone @xcite . in a simple participant spectator model , @xmath44 , where r is the radius of the nucleus at rest , @xmath45 is the spectator velocity in c.m .
and @xmath46 the corresponding lorentz factor . due to the comparable value of the passing time and expansion time in the collective expansion region up to 400 mev / nucleon
, the elliptical flow results an interplay between fireball expansion and spectator shadowing . in other words , due to the comparable size of two times(the fireball expansion and spectator shadowing ) , the participant particles which will come in the way of spectator shadowing are deflected towards the out - of - plane and hence more squeeze out .
however , in the energy range from 400 mev / nucleon to 1.49 gev / nucleon , @xmath43 decreases from 30 to 16 fm / c ( not shown here ) , implying that overall the expansion gets about two times faster in this energy region @xcite .
this is supported by the average expansion velocities extracted from the particle spectra @xcite . in this case , corresponding to much shorter passing time compared to the expansion time , the participant zone is affected by the shadowing of the spectator at very early times , however no shadowing effect is observed at later time where expansion of the participant matter is still happening . in other words , the participant particles at later times
are not blocked by the shadowing of spectator and hence decrease is observed in the out - of - plane emission after 400 mev / nucleon .
+ on the other hand , with increase in the rapidity region , the transition energy is found to affect to a great extent . in literature
, the balance energy using directed flow is also calculated , but was over the entire rapidity distribution @xcite .
the rapidity distribution affects many phenomena in intermediate energy region like particle production , nuclear stopping @xcite and now the elliptical flow .
the transition energy increases with rapidity region between @xmath47 and @xmath48 . with the increase in the rapidity region , dominance of the spectator matter from projectile as well as
target takes place that will further result in the dominance of the mean field up to higher energies . after the transition energy , the collective expansion
is found to have less squeeze out with an increase in the rapidity region . in this region ,
the spectator zone contribution along with participant zone starts to come in play .
as we know , the passing time for the spectator is very less compared to expansion time of the participant zone , leading to the decreasing effect of the spectator shadowing on the participant zone .
the chances of the participant to move in - plane increases with increase in the rapidity bin and hence less squeeze out is observed with increase in the size of the rapidity bin . if one see carefully , no transition is observed after @xmath49 .
this is due to negligible effect of the participant zone compared to the spectator zone .
the inset in the figure shows interesting results : the intersection of all the rapidity bins takes place at a particular incident energy . below and above this energy , incident energy dependence of elliptical flow
is changing the behavior with rapidity distribution .
one can have a good study of elliptical flow below and above this particular incident energy with variation in the rapidity distribution .
+ same as in fig.[fig:2 ] , but the contribution for fragments is from the projectile like matter only .
all the bins includes mid - rapidity region .
] same as in fig.[fig:2 ] , but the contribution for fragments is from the target like matter only .
all the bins includes mid - rapidity region . ] in fig.[fig:2 ] , we had displayed the effect on the participant as well as spectator matter from the projectile as well as target collectively including the distribution at @xmath39 .
it becomes quite interesting to study the participant and spectator matter contribution on the incident energy dependence of elliptical flow for projectile as well as target matter separately .
for this , in figs.[fig:3 ] and [ fig:4 ] , the incident energy dependence of the elliptical flow are displayed from the mid - rapidity to spectator zone for projectile and target matter .
both of the figures are following the universal behavior of the in - plane to out - of - plane emission with increase in incident energy , as is displayed in fig.[fig:2 ] .
on the other hand , the effect of rapidity bins is also quite similar as seen in of fig.[fig:2 ] . if one observes the maximum squeeze out values in figs .
[ fig:2]-[fig:4 ] around 400 mev / nucleon , it is found that less squeeze out is observed for projectile matter and spectator matter , separately as compared to projectile and target matter , collectively .
however , it is almost same for the projectile or target matter , separately .
this is obvious as when rapidity bin is from 0 to 0.1 or -0.1 to 0 , then participant zone is less compared to the region -0.1 to 0.1 .
this can be further clarify from the fig.[fig:1 ] , where rapidity distribution for different bins is displayed .
this is true for all the bins under investigation .
the effect of rapidity distribution on the transition energy will be discussed later under different conditions . + same as in fig.[fig:2 ] , but the contribution for fragments is from the projectile like matter only .
all the bins excludes the mid - rapidity region .
in this case , we move from participant+spectator contribution towards spectator contribution . ]
same as in fig.[fig:2 ] , but the contribution for fragments is from the target like matter only .
all the bins excludes the mid - rapidity region . in this case
, we move from participant+spectator contribution towards spectator contribution . ]
one notices from the above figures that major contribution for elliptical flow comes from the mid rapidity region .
one is further interested to know the fall of elliptical flow if the mid - rapidity region is excluded .
for this , by excluding the @xmath39 region , we have displayed the incident energy dependence for projectile and target matter from participant to purely projectile or target spectator matter in fig.[fig:5 ] and fig.[fig:6 ] , respectively .
the noted behavior is entirely different compared to previous three figures .
the differences are : ( a ) the behavior of rise and fall in the elliptical flow at a particular incident energy is obtained as were in the previous figures , but , no transition is observed from in - plane to out - of - plane emission . over all the incident energies , the value of elliptical flow remains positive , however , some competition is observed at low incident energies around 150 mev / nucleon .
this is due to the dominance of attractive mean field from the spectator matter , which restricts the effects of participant zone .
+ ( b ) a higher value of in - plane flow is observed at high incident energies compared to low incident energies upto @xmath50 in fig.[fig:5 ] for projectile matter and @xmath51 for target matter , which is in contrast to the previous figures .
this is due to the effect that at low incident energies , the passing time of spectator and expansion time of the participant and comparable , while , with increase in the incident energy , the passing time is found to be decrease as compared to expansion time .
this will reduce the shadowing effect at higher incident energies on the particles of the participant zone , results in the enhanced in plane flow .
moreover , more higher value of in - plane flow at higher incident energies as compared to lower one is due to the absence of most compressible region from the region @xmath52 , which competes with the mean field of the spectator to a great extent .
+ ( c ) with increase in the rapidity region from participant and spectator ( @xmath53 ) to purely spectator matter ( @xmath54 ) , the dominance of in - plane flow takes place upto @xmath50 and after that suddenly fall in the in - plane flow takes place upto @xmath54 in fig.[fig:5 ] , which is also true for the fig.[fig:6 ] .
the fall in the in - plane flow after @xmath50 can be explained with the help of the fig.[fig:1 ] .
after this region , the contribution of the participant is almost negligible and size of the spectator also decreases . due to the decreasing size of the spectator matter , the violence of the incident energy as well as short passing time of the spectator , they start to fly out - of - plane , which were enjoying in - plane earlier due to their heavier size .
hence , minimum in - plane flow is observed for @xmath54 where only 5 - 7 particles exist as is clear from the fig.[fig:1 ] . the same behavior is observed in the fig.[fig:6 ] . + the dependence of rapidity distributions on the transition energy .
the transition energies are extracted from fig.[fig:2 ] , [ fig:3]and [ fig:4 ] for top , middle and bottom panels , respectively .
the numbering from 1to 7 is representing the different rapidity bin from the respective figures mentioned above .
all the panels are parameterized with the straight line interpolation @xmath55 , where m is the slope , displayed in respective panels . ]
going through all the aspects of rapidity distribution , it is important to include the mid - rapidity region @xmath39 in the study of elliptical flow in heavy - ion collisions . in order to compare the findings with the experimental one
, one must have the transition from in - plane to out - of - plane , which are observed in figs.[fig:2]-[fig:4 ] .
moreover , more negative values are observed in fig.[fig:2 ] compared to other two one . from this discussion , it is fruitful to compare the findings of fig.[fig:2 ] with experimental findings of different collaborations .
this is displayed in fig.[fig:7 ] for lcp s ( z @xmath26 2 ) and protons .
the curves have the similar trend as displayed in fig.[fig:2 ] .
interestingly , it is observed that there is a competition in the rapidity bin @xmath56 and @xmath57 .
after that systematic deviation is observed in the elliptical flow values from the data values with increase in the rapidity bin .
there are some deviation in the middle region in case of lcp s between theory and data , while data is well explained for the protons . from here
, it is concluded that one can vary the mid rapidity region from the @xmath56 to @xmath57 to get the better agreement with the experimental one .
however , the discrepancy with the data for lcp s can be reduced by varying the isospin - dependent cross sections .
+ last , but not least , the rapidity distribution dependence of the transition energy is displayed in the fig.[fig:8 ] .
the top , middle and bottom panels are representing the transition energies extracted from fig.[fig:2]-[fig:4 ] , respectively .
the numbering from 1 to 7 in the figure is representing the bin size from the respective figures .
all the curves are fitted with straight line equation @xmath58 , where m is slope of line and c is a constant .
the transition energy is found to be sensitive towards the different bins of rapidity distributions .
it is observed that transition energy is found to increase with the size of the rapidity bin for light charged particles .
the necessary condition for the transition energy is that the mid - rapidity region must be included in the rapidity distribution bin .
it is also observed that no transition energy is obtained when the rapidity distribution region extends away from @xmath59 .
this is due to the dominance of the spectator matter , which enjoys in - plane as compared to out - of - plane .
the transition energy is found to be weakly sensitive towards the choice of the different type of rapidity distributions ( projectile - target or projectile or target ) while we have included the mid - rapidity region .
however , transition energy is found to disappear when mid - rapidity region is excluded . in other studies , system size dependence of the transition energy
is also studied by us and others for different kind of fragments @xcite .
recently , we have studied the fragment size dependence of transition energy under the influence of different equations of state , nucleon - nucleon cross sections etc @xcite .
these studies were made at a fixed rapidity bin .
the detailed analysis of the transition energy with rapidity distributions has revealed many interesting aspects for the first time . + for more interest , in the near future
, we are performing the comparative study of different qmd and iqmd simulations . from the preliminary results
, it is observed that isospin content of the colliding partners @xmath60 is supposed to playing the appreciable role in the analysis of elliptical flow in intermediate energy heavy - ion collisions in term of ( i ) closeness of the results with experimental findings of different collaborations and ( ii ) effect on the elliptical flow with change in the rapidity bins .
the detailed analysis of these findings will be presented in the near future .
within the semi classical transport simulations of energetic semi central collisions of @xmath8 reaction , we have carried out a new investigation of the interplay between the participant and spectator regions in term of rapidity distributions .
the maxima and minima in the incident energy dependence of elliptical flow is produced due to the different contributions of passing time of the spectator and expansion time of the participant .
the shadowing of the spectator matter plays important role up to later times due to the comparable magnitude of the passing and expansion time up to energy 400 mev / nucleon , however at high energies , shadowing effect is dominant only at earlier times due to the shorter passing time as compared to expansion time .
the transition from in - plane to out - of - plane is observed only when the mid - rapidity region is included in the rapidity bin , otherwise , no transition is observed .
the transition energy is found to be strongly dependent on the size of the rapidity bin , while , weakly dependent on the type of the rapidity distributions .
the transition energy is parameterized with a straight line interpolation .
comparison with experimental bin , reveals the competition is observed between the rapidity bin of @xmath47 and @xmath61 . to remove this discrepancy in the middle region for lcp s ,
one has to reduce the strength of nucleon - nucleon cross - section .
+ s. kumar , s. kumar and r. k. puri , phys .
c * 81 * , 014601 ( 2010 ) ; g. lehaut _
et al . , _ phys .
104 * , 232701 ( 2010 ) ; s. kumar and s. kumar , chin .
* 27 * , 062504 ( 2010 ) .
r. k. puri , s. s. malik and r. k. gupta , eur .
lett . * 9 * , 767(1989 ) ; r. k. puri and r. k. gupta , j. phys .
g : nucl . part
* 18 * , 903(1992 ) ; r. k. gupta , m. balasubramanium , r. k. puri and w. schied , j. phys .
g : nucl . part .
* 26 * , l23(2000 ) .
r. k. puri and r. k. gupta , phys .
c * 45 * , 1837(1992 ) ; r. k. puri , p. chattopadhyay and r.k .
gupta , phys .
c * 43 * , 315(1991 ) .
a. le fevre and j. aichelin , phys .
lett . * 100 * , 042701 ( 2008 ) ; a. le fevre _
et al_,phys .
rev . * c 80 * , 044615(2009 ) ; y. k. vermani , j. k. dhawan , s. goyal , r. k. puri and j. aichelin , j. phys .
g : nucl . part .
* 37 * , 015105(2010 ) .
g. d. westfall _ et al .
lett . * 71 * , 1986 ( 1993 ) ; m. b. tsang _
et al . , _ phys .
c * 53 * , 1959 ( 1996 ) ; y. m. zheng , c. m. ko , b. a. li , and b. zhang , phys . rev .
lett . * 83 * , 2534 ( 1999 ) ; a. b. larionov , w. cassing , c. greiner , and u. mosel , phys .
c * 62 * , 064611 ( 2000 ) ; b. a. li , a. t. sustich , and b. zhang , _ ibid . _ * 64 * , 054604 ( 2001 ) ; c. alt _ et al . , _ _ ibid._*68 * , 034903 ( 2003 ) . a. d. sood and r. k. puri , phys . rev .
c * 69 * , 054612 ( 2004 ) ; eur .
a * 30 * , 571 ( 2006 ) ; a. d. sood and r. k. puri , phys .
c * 70 * , 034611 ( 2004 ) ; s. kumar , m. k. sharma , r. k. puri , k. p. singh , and i. m. govil , _ ibid . _ * 58 * , 3494 ( 1998 ) , a. d. sood , r. k. puri , and j. aichelin , phys . lett .
b * 594 * , 260 ( 2004).s .
kumar and r. k. puri , phys .
c * 60 * , 054607 ( 1999 ) ; s. kumar , s. kumar , and r. k. puri , phys .
c * 78 * , 064602 ( 2008 ) ; y. k. vermani , s. goyal , and r. k. puri , _
ibid._*79 * , 064613 ( 2009 ) ; r. chugh and r. k. puri , phys . rev .
* c 82 * , 014603 ( 2010 ) .
et al . , _ phys .
76 * , 3911 ( 1996 ) ; d. brill _ et al . , _ z. phys .
a * 355 * , 61 ( 1996 ) ; m.b .
et al . , _ phys .
c * 53 * , 1959 ( 1996 ) ; n. bastid _ et al .
a * 622 * , 573 ( 1997 ) ; p. crochet _
et al . , _ nucl .
a * 624 * , 755 ( 1997 ) ; g. stoicea _
et al . , _ phys . rev .
lett . * 92 * , 072303 ( 2004 ) .
d. j. magestro , w. bauer , and g. d. westfall , phys .
c * 62 * , 041603(r ) ( 2000 ) ; j. singh and r. k. puri , phys .
c * 62 * , 054602 ( 2000 ) ; a. d. sood and r. k. puri , phys .
c * 73 * , 067602 ( 2006 ) .
r. k. puri , ch . hartnack and j. aichelin ,
c * 54 * , r28 ( 1996 ) .
p. b. gossiaux , r. k. puri , ch . hartnack and j. aichelin ,
a * 619 * , 379(1997 ) .
r. k. puri and j. aichelin , j. comp .
phys . * 162 * , 245(2000 ) . | interplay between the spectator and participant matter in heavy - ion collisions is investigated within isospin dependent quantum molecular dynamics ( iqmd ) model in term of rapidity distribution of light charged particles .
the effect of different type and size rapidity distributions is studied in elliptical flow .
the elliptical flow patterns show important role of the nearby spectator matter on the participant zone .
this role is further explained on the basis of passing time of the spectator and expansion time of the participant zone .
the transition from the in - plane to out - of - plane is observed only when the mid - rapidity region is included in the rapidity bin , otherwise no transition occurs .
the transition energy is found to be highly sensitive towards the size of the rapidity bin , while weakly on the type of the rapidity distribution .
the theoretical results are also compared with the experimental findings and are found in good agreement . |
particles which contain strange quarks are valuable probes of the dynamics of @xmath4 collisions as constituent strange quarks are not present in the initial colliding nuclei .
the enhancement of the strange particle yield from @xmath4 collisions to heavy ion collisions has been suggested as a possible quark gluon plasma signature @xcite .
+ we report preliminary results from the 2001/2002 p+p run by the star experiment .
the main focus of this paper is on presenting a high statistics measurement of @xmath1 , @xmath2 and @xmath3 at @xmath5 gev and obtaining the yield and @xmath0 for each species .
a discussion of appropriate parameterizations to the particle spectra will ensue .
we will show that the widely used power - law extrapolation in @xmath6 collisions is not sufficient to obtain the best @xmath7 results for fits to the strange particle spectra and we will consider alternatives@xcite .
+ furthermore , we will study the dependency of @xmath8 and spectral shape as a function of charged particle event multiplicity(@xmath9 ) , an effect that has been observed previously in high energy @xmath6 collisions .
several authors ascribe this phenomenon to pqcd mini - jets fragmenting into hadronic final states @xcite .
however , we show that present string models incorporating semi - hard pqcd processes , such as pythia , do not describe the star data well without significant parameter tuning @xcite .
the data were reconstructed using the star detector system which is described in more detail elsewhere @xcite .
the main tracking detector used in this analysis is the time projection chamber ( tpc ) covering the full acceptance in azimuth and a large pseudo - rapidity coverage ( @xmath10 ) . +
a total of 14 million non - singly diffractive ( nsd ) events were triggered with the star beam - beam counters ( bbc ) requiring two coincident charged tracks at forward rapidity ( @xmath11 ) . due to the particulary low track multiplicity environment in p+p collisions
a special vertex finder was developed .
a monte - carlo study showed that only 76% of primary vertices are found correctly ; from the remainder , 14% are lost and 10% are badly reconstructed .
a total of 11 million events passed the event level selection criteria requiring a valid primary vertex within 100 cm along the beam - line from the center of the tpc detector .
an additional multiplicity dependant vertex efficiency factor @xmath12 must therefore be applied to the particle spectra .
+ the high luminosity of the rhic proton beams has the unfortunate effect of producing several collisions during the read - out time of the tpc , a condition known as pile - up " .
these pile - up events can not be identified separately , however the triggered vertex can be identified by allowing only tracks that match to a very fast scintillating central trigger barrel(ctb ) detector to be used in the vertex reconstruction .
the strange particles were identified from their weak decay to charged daughter particles via topological identification algorithm .
the following decay channels were analyzed : @xmath13(b.r .
63.9% ) , @xmath14(b.r .
63.9% ) , @xmath15 ( b.r .
+ particle identification of daughters was achieved by requiring the de / dx to fall within the 3@xmath16-bands of the theoretical bethe - bloch parameterizations .
further background in the invariant mass was removed by applying topological cuts to the decay geometry , _ i.e. _ requiring the parent to originate from within 2 cm of the primary vertex and have a minimum decay length of 2 cm .
these cuts also ensure that the particle decay is matched to the correct vertex with a high probability .
+ corrections for particle reconstruction efficiency were obtained through a monte - carlo based method of embedding simulated particle decays into real events and comparing the number of simulated and reconstructed particles in each @xmath17-bin .
the @xmath2 and @xmath3 spectra were corrected for feed - down from @xmath18 ( @xmath19 ) . finally , the corrected transverse momentum spectra have to be multiplied by the multiplicity dependant primary vertex efficiency @xmath12 to account for yield losses .
non - feeddown corrected spectra for @xmath1 , @xmath2 and @xmath3 are shown in figure [ fig : allspec ] .
the particle acceptance at mid - rapidity ( @xmath20 ) in the tpc starts at transverse momentum 0.2 gev / c for @xmath1 and 0.3 gev / c for @xmath2 and @xmath3 .
the spectra are parameterized to allow extrapolation to low pt and to extract @xmath0 and yield ( @xmath21 ) .
in contrast to previous @xmath6 experiments @xcite , which used either a single exponential function in transverse mass or power - law functions , we found that a combination of these functions is more effective in fitting the star data .
the @xmath7 results from different fit - functions are presented in table [ tab : fits1 ] .
composite fits , represented in equation 1 ( for @xmath1 ) and equation 2 ( for @xmath2 and @xmath3 ) yield the lowest @xmath7 and were used to extract the final values for @xmath22 and yield as shown in table [ tab : results1 ] .
.a summary of @xmath7 values for different fit - functions to the particle spectra [ cols="^,^,^,^,^,^,^,^",options="header " , ]
figure [ fig : spectrapythia ] shows the comparison of the star spectra to the spectra from the default setting of pythia event generator ( version 6.22 , msel1 ) which includes hard processes and a modelling of soft processes , through a pqcd model . even at @xmath23 1.5 ,
a regime were pqcd should give a better description , the differences between the model and the data are considerably large .
further studies are planned in order to optimally tune the model to match the star strangeness data .
the minimum bias event sample was split into 6 event classes according to mean charged particle multiplicity per unit @xmath24 ( @xmath25 ) .
a similar analysis has been performed in previous @xmath6 experiments at different collision energies @xcite .
high multiplicity events are particularly interesting as they are likely dominated by contributions from large momentum transfer parton - parton collisions .
figure [ fig : meanptpythia ] presents a comparison of the @xmath22 vs @xmath25 for @xmath2 and @xmath1 with the corresponding pythia calculations . a rise in @xmath0 with increasing @xmath9
is observed and is stronger for the heavier particle ( @xmath2 ) .
several authors have attributed these phenomena to the increased number of fragmenting mini - jets in the high - multiplicity events @xcite .
in fact , the hijing model even makes a prediction for the expected number of mini - jets contributing to each multiplicity @xcite .
it is also interesting that pythia under - predicts the magnitude and correlation strength of @xmath0 with multiplicity .
similar comparisons for inclusive charged particles @xmath0 have been recently published by the cdf collaboration and demonstrate similar discrepancies @xcite .
in figure [ fig : meanptmass ] the @xmath0 is shown for all particle species measured in @xmath4 collisions in star as a function of particle mass .
the black line is an empirical parameterization to isr data for @xmath26,k , p only , at @xmath27 = 25 gev .
there is very little energy dependance for the lower mass particles between the isr and rhic energies , whereas the higher mass particles show a clear breaking from the parameterization .
the dependance of @xmath0 on particle mass has been explained by mini - jet production in @xmath4 and @xmath6 collisions and a stronger contribution for higher mass particles is expected @xcite .
the star experiment has made the first high statistics measurement of mid - rapidity @xmath1 , @xmath2 and @xmath3 in @xmath4 collisions at @xmath27 = 200 gev .
the results agree with those made by the ua5 collaboration for @xmath6 at @xmath27 = 200 gev .
the ratio of @xmath3/@xmath2 suggests a small net baryon number at mid - rapidity .
we have show that the shape of the particle spectra is best described with a two - component fit function accounting for both the soft and hard parts of the spectra .
furthermore , we have undertaken studies to understand the change in shape and @xmath0 of the spectra with increased event multiplicity .
we compared our results to pqcd inspired models which infer mini - jets and string fragmentation .
we have shown that the current default version of pythia is not well tuned for strange particles at this collision energy .
99 rafelski j and mller b _ phys . rev .
lett . _ * 48 * ( 1982 ) 1066 bocquet g et al ( ua1 collaboration ) _ phys .
_ b * 366 * ( 1996 ) 441 sjstrand t and van zijl m _ phys .
_ d vol * 36 * ( 1987 ) 2019 wang x and gyulassy m _ phys .
_ b * 282 * ( 1992 ) 466 sjstrand t et al ( pythia 6.2 physics and manual ) , _ preprint _ hep - ph/0308153 ( 2002 ) ackermann k h et al ( star collaboration ) _ nucl .
_ a*661 * ( 1999 ) 681 ackermann k h et al ( star collaboration ) _ nucl .
_ a*499 * ( 2003 ) 624 ansorge r e et al ( ua5 collaboration ) _ nucl .
_ b*328 * ( 1989 ) 36 alexopolous t et al ( e735 collaboration ) _ phys .
_ d vol * 48 * ( 1993 ) 984 wang x and gyulassy m _ phys .
_ d * 45 * ( 1992 ) 844 gaisser t k and halzen f _ phys .
* 54 * ( 1985 ) 1754 pancheri g and srivastava y
_ b * 182 * ( 1986 ) 199 acosta d et al ( cdf collaboration ) _ phys .
_ d * 65 * ( 2002 ) 072005 | we present measurements of the transverse momentum spectra , yield and @xmath0 systematics for @xmath1 , @xmath2 and @xmath3 in @xmath4 collisions at @xmath5 gev .
we show a dependence of the @xmath0 with event multiplicity and infer that this is consistent with a mini - jet dominated particle production mechanism .
these observations are compared to available data from @xmath6 experiments as well as to pqcd theoretical predictions . |
dynamical breaking of a symmetry can be described by a quantity known as the order parameter . in the case of a broken chiral symmetry
a commonly used order parameter is the chiral condensate which vanishes if the chiral symmetry is unbroken .
the investigation of the chiral condensate plays a crucial role in attempts to describe phase transitions related to the dynamical chiral symmetry breaking . in qcd chiral condensate
is obtained by numerical studies , as a rule via lattice simulations .
this approach is usually employed in studies of the chiral condensate in the presence of external factors such as temperature , chemical potentials , magnetic field , finite size effects , etc ( see [ 1 - 3 ] and references therein ) . apart from numerical studies there is a lot of models employed for the investigation of the chiral condensate in various physical systems @xcite , @xcite . in the first paper on this subject @xcite nambu and
jona - lasinio ( njl ) analyzed a specific field model in four dimensions .
later , the chiral condensate was studied by gross and neveu ( gn ) in two - dimensional spacetime in the limit of a large number of fermion flavors @xmath0 @xcite .
these two models are similar but in contrast to the njl model , the gn model is a renormalizable and asymptotically free theory . due to these properties
the gn model is used for qualitative modeling of qcd .
the relative simplicity of both models is a consequence of the quartic fermion interaction . in our previous paper
@xcite we investigated a system of a self - interacting massive scalar and a massless fermion field with the yukawa interaction in a @xmath1-dimensional spacetime . in the limit of a large mass of the scalar field
the model becomes equivalent to the gn model .
the chiral condensate was obtained by the functional integration using the method of the saddle point . for technical reasons the contribution of the scalar field fluctuations was not taken into account .
this can be done if the coupling constant of the self - interacting scalar field is much less than the yukawa coupling constant .
continuing our investigation of this model , we present in this paper a study of how the chiral condensate is affected by the quantum fluctuations of the scalar fields .
the consideration of the quantum fluctuations allows one to find a phase transition in this model in a @xmath1-dimensional spacetime .
apart from that , we also investigate the chiral condensate in a @xmath2-dimensional spacetime . in the limit of the massless scalar field in @xmath2 dimensions ,
the chiral condensate is obtained in analytic form .
we also find a phase transition in this setting .
all results are obtained in the one - loop approximation framework .
we consider a model with the lagrangian density given by @xmath3 where @xmath4 is a real scalar field , @xmath5 is a massless fermion field and the index @xmath6 runs from 1 to @xmath0 .
the potential of the scalar field @xmath7 includes the mass term and the self - interaction @xmath8 of the scalar field .
the potential @xmath9 is the yukawa interaction . @xmath10 where @xmath11 is a coupling constant .
the lagrangian ( [ eq1 ] ) is invariant under the discrete transformation @xmath12 the symmetry ( [ eq3 ] ) is broken by the chiral condensate . in this paper
we extend the calculation of the chiral condensate in this model performed in ref .
@xcite to include the quantum fluctuations of the scalar field .
we formally define the chiral condensate using the functional integral @xmath13 where @xmath14 is a normalization constant .
( [ eq4 ] ) can be rewritten as @xmath15 @xmath16 to integrate over the scalar field one decomposes @xmath17 , where @xmath18 satisfies the classical equation of motion while @xmath19 describes small fluctuations around the classical background . in order to study the properties of the chiral condensate we choose a specific potential of the scalar field as @xmath20 where @xmath21 is a coupling constant .
then we take into account the quadratic terms of the scalar fluctuation @xmath19 and integrate over them , getting @xmath22 where @xmath23 . integrating over the fermionic field , we obtain @xmath24 where @xmath25 is a fermionic ultraviolet cutoff , @xmath26 is a ground state value of the bosonic field which is determined by the equations @xmath27 where @xmath28 , whereas @xmath11 and @xmath21 are dimensional coupling constants .
it is convenient to introduce dimensionless constants @xmath29 where @xmath30 is a bosonic ultraviolet cutoff .
then eq .
( [ eq7 ] ) may be written as @xmath31 and the saddle point @xmath26 is given by the solution of the equation @xmath32 where all the quantities are dimensionless .
the third term in this equation describes the contribution of the scalar field fluctuations . generally speaking , eq .
( [ eq9 ] ) is resolved numerically , however , some special cases lead to simplifications that allow one to solve it analytically .
an analytic solution can be easily compared with the results of our previous work @xcite , where the quantum fluctuations of the scalar field were not accounted for .
we consider two such possibilities .
\i ) when @xmath33 , then @xmath34 and substituting @xmath26 in ( [ eq8 ] ) , we get @xmath35 to see that this solution coincides with that of ref .
@xcite , we note that @xmath36 equals @xmath26 of that work ( see eq .
( 15 ) there ) .
\ii ) when @xmath37 , then , rescaling @xmath38 in eq .
( [ eq9 ] ) , we get @xmath39 where @xmath40 using ( [ eq10 ] ) and ( [ eq8 ] ) , we obtain @xmath41 if @xmath42 , then the value of the chiral condensate given by eq .
( [ eq11 ] ) equals that of eq .
( 13 ) in ref .
@xcite .
the value of @xmath43 is renormalization invariant .
indeed , the @xmath44-functions in the one - loop approximation are @xmath45 using ( [ eq12 ] ) one can verify that @xmath43 does not depend on the choice of the renormalization point , i.e. @xmath46 it is worth to note that in a @xmath1-dimensional spacetime there is only one @xmath14-factor that appears in the renormalization of the scalar field mass , therefore @xmath47 where @xmath48 and @xmath49 are the bare coupling constants , and @xmath50 , @xmath51 are the renormalized coupling constants .
the relation in eq .
( 13 ) can be obtained directly from the @xmath44-functions of eq .
( [ eq12 ] ) . using eq .
( [ eq13 ] ) @xmath43 is rewritten as @xmath52 one can see that @xmath43 and the @xmath44-functions tend to zero in the limit @xmath53 . on the other hand ,
if the @xmath44-function equals to zero then in the model there is a phase transition as shown in ref .
therefore one can conclude that if @xmath54 is @xmath55 then there is a phase transition in the model .
although @xmath43 is not a solution of the eq .
( [ eq9 ] ) or eq .
( [ eq10 ] ) , @xmath43 also tends to zero when the @xmath44-functions tend to zero therefore @xmath43 may be considered as a qualitative approximation of @xmath26 .
the value of @xmath43 becomes a good enough approximation of @xmath26 provided @xmath56 .
the results of our numerical studies of the chiral condensate are presented in fig .
1 . for the convenience
the value @xmath57 is shown in this figure .
we plot the value @xmath58 as a function of the yukawa coupling constant @xmath59 .
the other parameters entering eq .
( [ eq9 ] ) take the values @xmath60 .
since the chiral condensate falls sharply when @xmath54 tends to 1 , then @xmath58 is presented in different scales in fig .
1 ( a , b ) . the solid line corresponds to @xmath26 obtained from the solution of eq .
( [ eq9 ] ) .
the dotted line corresponds to @xmath26 obtained without the contribution of the scalar field fluctuations .
the dash - dotted line is obtained for @xmath61 .
we see that the chiral condensate equals to zero and the chiral symmetry is unbroken provided @xmath54 is @xmath55 .
the chiral condensate appears when this condition is violated and increases with @xmath48 .
hence there is a phase transition in this model . with increasing @xmath48
the quantity @xmath43 becomes a good approximation of @xmath26 according to the corresponding condition @xmath56 . as a function of the yukawa coupling constant @xmath59
the values of parameters are @xmath60 .
the solid line shows the chiral condensate where @xmath26 is an exact solution of eq .
( [ eq9 ] ) .
the dotted line is @xmath58 obtained without the contribution of the scalar field fluctuations .
the dash - dotted line is @xmath58 obtained for @xmath61 . ]
the method used above to calculate the chiral condensate is applicable in any dimension of space . therefore calculating @xmath62
we get @xmath63 @xmath64 the quadratically divergent terms which appear in @xmath65 are eliminated by the mass renormalization .
further we consider the case of a massless scalar field @xmath66 . rescaling the ultraviolet cutoff @xmath67 we rewrite the effective potential ( [ eq15 ] ) as @xmath68 or @xmath69 where @xmath70 note that @xmath43 is not a renormalization - invariant quantity .
it is possible to check that by direct calculation , but it can be done in a simpler way .
namely , we consider the following limiting cases \i ) @xmath71 , @xmath72 , \ii ) @xmath73 , @xmath74 is not a renormalization - invariant quantity , we recall that @xmath75 . in four dimensions the chiral condensate is defined as @xmath76 where @xmath26 is determined by the equations @xmath77 the effective potential @xmath78 has a minimum at @xmath79 and @xmath80 . in this case the chiral condensate is given by @xmath81 in the limit @xmath82 , @xmath43 tends to zero and consequently @xmath83 .
if @xmath84 then @xmath62 has a maximum , the chiral condensate does not exist , and the chiral symmetry is unbroken .
in this paper we studied the dynamical symmetry breaking in models consisting of a fermion field and a self - interacting scalar field with the yukawa interaction between the two fields in the one - loop approximation .
we believe that the dynamical symmetry breaking as exhibited by these models can help to analyze similar phenomena in more realistic models .
the models are considered in two and four spacetime dimensions .
we showed that each models has two phases , one with the preserved symmetry and the other with the broken symmetry .
we conclude that both models have a phase transitions as a function of the coupling constants . in four spacetime dimensions ,
we obtain an analytic expression for the chiral condensate in the limit of a massless scalar field .
the two - dimensional case was investigated numerically by solving of the algebraic equation for the scalar field .
buividovich , m.n .
chernodub , e.v .
luschevskaya and m.i .
polikarpov , phys . lett . *
b682 * , 484 ( 2010 ) ; v.v .
braguta , p.v .
buividovich , t. kalaydzhyan , s.v .
kuznetsov , m.i .
polikarpov , arxiv : 1011.3795v3 [ hep - lat ] .
d. ebert , n.v .
gubina , k.g .
klimenko , s.g .
kurbanov and v. ch .
zhukovsky , phys . rev . *
d84 * , 025004 ( 2011 ) ; d. ebert .
khunjua , k.g .
klimenko and v. ch .
zhukovsky , arxiv : 1306.4485v2 [ hep - th ] .
g. thooft , nucl . phys . *
b75 * , 461 ( 1974 ) ; a. zhitnitsky , phys . rev . * d53 * , 5821 ( 1996 ) ; l.ya .
glozman , v.k .
sazonov , m. shifman , r.f .
wagenbrunn , arxiv : 1201.5814v1 [ hep - th ] ; s. bellucci , e.r .
bezerra de mello , a.a .
saharian , arxiv : 1312.1686v1[hep - th ] . y. nambu and g. jona - lasinio , phys . rev . * 122 * , 345 ( 1961 ) .
gross and a. neveu , phys .
rev . * d10 * , 3235 ( 1974 ) .
v.g . ksenzov and a.i .
romanov , phys .
nucl . * 76 * , 534 ( 2013 ) ; arxiv : 1204.3051v2 [ hep - th ] .
a.m. polyakov , pisma zh .
* 12 * , 381 ( 1970 ) . | we investigate a class of models with a massless fermion and a self - interacting scalar field with the yukawa interaction between these two fields .
the models considered are formulated in two and four spacetime dimensions and possess a discrete symmetry .
we calculate the chiral condensates are calculated in the one - loop approximation .
we show that the models have a phase transitions as a function of the coupling constants . |
several criteria to detect non - classicality of quantum states of a harmonic oscillator have been introduced , mostly based on phase - space distributions @xcite , ordered moments @xcite , or on information - theoretic arguments @xcite . at the same time
, an ongoing research line addresses the characterization of quantum states according to their gaussian or non - gaussian character @xcite , and a question arises on whether those two different hierarchies are somehow linked each other . as a matter of fact ,
if we restrict our attention to pure states , hudson s theorem @xcite establishes that the border between gaussian and non - gaussian states coincides exactly with the one between states with positive and negative wigner functions .
however , if we move to mixed states , the situation gets more involved .
attempts to extend hudson s theorem have been made , by looking at upper bounds on non - gaussianity measures for mixed states having positive wigner function @xcite . in this framework , by focusing on states with positive wigner function , one can define an additional border between states in the _ gaussian convex hull _ and those in the complementary set of _ quantum non - gaussian states _
, that is , states that can not be expressed as mixtures of gaussian states .
the situation is summarized in fig .
[ f : poswigner ] : the definition of the gaussian convex hull generalizes the notion of glauber s non - classicality @xcite , with coherent states replaced by generic pure gaussian states , i.e. squeezed coherent states
. quantum non - gaussian states with positive wigner function are not useful for quantum computation @xcite , and are not necessary for entanglement distillation , _
e.g. _ the non - gaussian entangled resources used in @xcite are mixtures of gaussian states . on the other hand , they are of fundamental interest for quantum information and quantum optics . in particular , since no negativity of the wigner function can be detected for optical losses higher than @xmath0 @xcite ( or equivalently , for detector efficiencies below @xmath0 ) criteria able to detect quantum non - gaussianity are needed in order to certify that a _ highly non linear _ process ( such as fock state generation , kerr interaction , photon addition / subtraction operations or conditional photon number detections ) has been implemented in a noisy environment , even if no negativity can be observed in the wigner function .
different measures of non - gaussianity for quantum states have been proposed @xcite , but these can not discriminate between quantum non - gaussian states and mixtures of gaussian states . an experimentally friendly criterion for quantum non - gaussianity , based on photon number probabilities , has been introduced @xcite , and then employed in different experimental settings to prove the generation of quantum non - gaussian states , such as heralded single - photon states @xcite , squeezed single - photon states @xcite and fock states from a semiconductor quantum dot @xcite . ] in this paper we introduce a family of criteria which are able to detect quantum non - gaussianity for single - mode quantum states of a harmonic oscillator based on the wigner function . as we already pointed out , according to hudson s theorem ,
the only pure states having a positive wigner function are gaussian states .
one can then wonder if any bound exists on the values that the wigner function of convex mixtures of gaussian states can take . by following this intuition we present several bounds on the values of the wigner function for convex mixtures of gaussian states , consequently defining a class of sufficient criteria for quantum non - gaussianity .
in the next section we will introduce some notation and the preliminary notions needed for the rest of the paper . in sec .
[ s : criteria ] we will prove and discuss our wigner function based criteria for quantum non - gaussianity and in sec . [
s : examples ] we will prove their effectiveness by considering different families of non - gaussian states evolving in a lossy ( gaussian ) channel . we will conclude the paper in sec . [
s : conclusions ] with some remarks .
throughout the paper we will use the quantum optical terminology , where excitations of a quantum harmonic oscillator are called photons .
all the results can be naturally applied to any bosonic continuous - variable ( cv ) system .
we will consider a single mode described by a mode operator @xmath1 , satisfying the commutation relation @xmath2 = \mathbbm{1}$ ] .
a quantum state @xmath3 is fully described by its characteristic function @xcite @xmath4(\gamma ) = { { \rm tr}}[\varrho d(\gamma)]\:,\end{aligned}\ ] ] where @xmath5 represents the displacement operator .
in addition , the quantum state @xmath3 can be fully described by the fourier transform of the characteristic function , _
i. e. _ the wigner function @xcite @xmath6(\alpha ) = \int \frac{d^2 \gamma}{\pi^2 } e^{\gamma^ * \alpha - \gamma \alpha^ * } \chi[\varrho](\gamma)\:.\end{aligned}\ ] ] a state is defined to be _ gaussian _ if and only if its wigner function ( or equivalently its characteristic function ) is gaussian .
all single - mode gaussian states can be expressed as @xmath7 where @xmath8 , is the squeezing operator , and @xmath9 $ ] is a thermal state ( @xmath10 and @xmath11 ) .
pure gaussian states can be written as @xmath12 , and , according to hudson s theorem @xcite , they are the only pure states having a positive wigner function . together with that of a gaussian state , one can define the concept of _ gaussian map _ : a quantum ( completely positive ) map is defined gaussian iff it transforms gaussian states into gaussian states .
all unitary gaussian maps can be expressed as @xmath13 , and they correspond to hamiltonian operators @xmath14 at most bilinear in the mode operators .
similarly , a generic gaussian map can be decomposed as a gaussian unitary acting on the system plus an ancilla ( the latter prepared in a gaussian state ) , followed by partial tracing over the ancillary mode @xcite .
another complete description of a cv quantum state @xmath3 may be given in terms of the so - called @xmath15-function @xmath16(\alpha)$ ] @xcite , defined implicitly via the formula @xmath17(\alpha ) where @xmath18 represents a coherent state .
according to glauber a state is _ non - classical _ iff its @xmath15-function is not a proper probability distribution , _
e.g. _ the @xmath15-function is more singular than a dirac - delta function .
note that the negativity of the wigner function is a more restrictive definition of non - classicality : there exists non - classical states having a positive wigner function ( _ e.g. _ squeezed states ) , while all the states having a non - positive wigner function are non - classical according to glauber . in a similar spirit as in glauber s approach to non - classicality , in this paper we study the concept of _ quantum non - gaussian _ states .
these are defined as follows .
the gaussian convex hull is the set of states @xmath19 where @xmath20 denotes the hilbert space of continuous - variable quantum states , @xmath21 is a proper probability distribution and @xmath22 are pure gaussian states , i.e. , in the single mode case , squeezed coherent states identified by the set of parameters @xmath23 . since gaussian states do not form a convex set , the set in eq
. includes states which are not gaussian .
moreover any mixed gaussian state can be written as a weighted sum of pure gaussian states , and
hence the set above also includes convex mixtures of mixed gaussian states .
the definition of _ quantum non - gaussianity _ naturally follows : * definition*. _ a quantum state @xmath3 is quantum non - gaussian iff it is not possible to express it as a convex mixture of gaussian states , that is iff @xmath24_. as illustrated in fig .
[ f : poswigner ] , the border here defined dividing quantum non - gaussian states and mixtures of gaussian states falls in between the border dividing classical and non - classical states , and the one which divides states with positive and non - positive wigner functions .
the importance of such a further distinction is evident if we note that all states in @xmath25 can be prepared through a combination of gaussian operations and classical randomization . on the contrary , if @xmath26 then a highly _ non - linear _ process ( due to a non - gaussian operation or measurement ) had necessarily taken place in the generation the quantum state @xmath3 . while the negativity of the wigner function is always sufficient to certify it , more elaborated criteria , as those elaborated in this paper ,
are needed in order to detect such a characteristic when quantum states exhibit a positive wigner function .
in order to find criteria for the detection of quantum non - gaussian states , we follow the intuition given by hudson s theorem for pure gaussian states . we will focus on lower bounds on the values taken by the wigner function of states which belong to the gaussian convex hull @xmath27 . in this section we present our main findings as one lemma leading to two final propositions and two additional corollaries .
the ` quantum non - gaussianity criteria ' , derived directly from these results , are presented at the end of the section .
[ l : boundpure ] for any given pure single - mode gaussian state @xmath28 , the value of the wigner function at the origin of the phase space is bounded from below as @xmath29(0 ) \geq \frac2\pi \exp\ { -2 n ( 1+n)\ } \ : , \label{eq : boundpure}\end{aligned}\ ] ] where @xmath30 . *
proof * a generic pure single - mode gaussian state can be always written as @xmath31 , where @xmath32 and @xmath33 ( @xmath34 ) are two complex numbers .
we can thus write the wigner function evaluated in zero as @xmath35&(0)=\frac{2}{\pi}\exp\left\{-2|\alpha|^2 \left[\cosh{2r } \right . \right.\nonumber \\ & \left.\left .
-\cos{(2\theta+\phi)}\sinh{2r } \right ] \right\ } .
\label{eq : wiggausszero}\end{aligned}\ ] ] our goal is to minimize the value of the wigner function or equivalently to maximize the function @xmath36 a first maximization is obtained by considering @xmath37 which yields @xmath38 in the last equation we introduced the displacement and squeezing photon numbers , @xmath39 and @xmath40 , and we used the formula @xmath41 .
note that these two parameters obey @xmath42 where @xmath43 is the average photon number of the state @xmath44 .
we can thus express the right hand side ( rhs ) of eq .
( [ eq : max1 ] ) in terms of @xmath43 and @xmath45 , obtaining @xmath46 for a given average photon number @xmath43 , the above function is maximized with regard to the parameter @xmath45 by choosing @xmath47 and obtaining @xmath48 this leads to @xmath49(0 ) \geq
\frac 2\pi \exp\left\{-2n(1+n)\right\ } .\end{aligned}\ ] ] @xmath50 by looking at the proof , we remark that the bound obtained is tight : given a fixed energy @xmath43 , by choosing the phases according to condition ( [ eq : phases ] ) and the squeezing energy according to ( [ eq : optsq ] ) , it is always possible to find a family of pure gaussian states saturating the inequality .
in particular the maximization obtained via condition ( [ eq : phases ] ) simply corresponds , at fixed @xmath51 and @xmath45 , to displace the state along the direction of the squeezed quadrature .
the condition ( [ eq : optsq ] ) shows that for small values of @xmath43 , the minimum of the wigner function is obtained by using the energy in displacement , while for larger values of @xmath43 , the optimal squeezing fraction @xmath45 tends to an asymptotic value @xmath52 .
+ let us now generalize the bound obtained to a generic convex mixture of gaussian states .
[ p : boundmix ] for any single - mode quantum state @xmath3 which belongs to the gaussian convex hull @xmath27 , the value of the wigner function at the origin is bounded by @xmath6(0 ) \geq \frac2\pi \exp\ { -2 \bar{n } ( 1+\bar{n})\ } \label{eq : boundmix}\end{aligned}\ ] ] where @xmath53 $ ] . * proof * the multi - index @xmath54 , which labels every gaussian state in the convex mixture @xmath55 , contains the information about the squeezing @xmath56 and displacement @xmath57 .
we can then equivalently consider as variables @xmath58 . by exploiting the linearity property of the wigner function
we obtain @xmath6(0 ) & = \int d{{\boldsymbol \lambda}}\ : p({{\boldsymbol \lambda } } ) w[|\psi_{\sf g}({{\boldsymbol \lambda}})\rangle\langle\psi_{\sf g}({{\boldsymbol
\lambda}})|](0 ) \nonumber \\ & \geq \frac 2\pi \int d{{\boldsymbol \lambda}}\ : p({{\boldsymbol \lambda } } ) \exp \ { -2n(1+n ) \ } \ : , \label{eq : integral}\end{aligned}\ ] ] where inequality ( [ eq : boundpure ] ) has been used . by defining @xmath59 which is a valid probability distribution with respect to the variable @xmath43 , eq .
( [ eq : integral ] ) becomes@xmath6(0 ) \geq \frac 2\pi \int_0^{\infty } dn \ : \widetilde{p}(n ) \ : \exp \ { -2n(1+n)\}.\end{aligned}\ ] ] studying the second derivative of @xmath60 we conclude that the function is _ convex _ in the whole _ physical _ region ( _ i.e. _ @xmath61 ) . as a consequence , @xmath62 where @xmath63 $ ] . from the last inequality
we obtain straightforwardly the thesis @xmath6(0 ) \geq
\frac 2\pi \exp\ { -2 \bar{n } ( 1+\bar{n } ) \ }
\:.\end{aligned}\ ] ] @xmath50 the following proposition generalizes the bound obtained above .
[ p : bound2mix ] for any single - mode quantum state @xmath64 , and for any given gaussian map @xmath65 ( or alternatively a convex mixture thereof ) , the following inequality holds @xmath66(0 ) \geq \frac2\pi
\exp\ { -2 \bar{n}_\mathcal{e } ( 1+\bar{n}_\mathcal{e})\}\ : , \label{eq : bound2mix}\end{aligned}\ ] ] where @xmath67 $ ] . * proof * given a quantum state @xmath3 which can be expressed as a mixture of gaussian state , and a gaussian map @xmath65 ( or a convex mixture thereof ) , the output state @xmath68 can still be expressed as a mixture of gaussian states . as a consequence we can apply to the state @xmath69 the result in proposition [ p : boundmix ] , obtaining the thesis .
@xmath50 + proposition [ p : bound2mix ] leads to two corollaries that will be used in the rest of the paper . for any single - mode quantum state @xmath64 , the following inequality holds @xmath6(\beta ) \geq \frac2\pi \exp\ { -2 \bar{n}_\beta ( 1+\bar{n}_\beta)\}\ : , \:\ : \forall \beta \in \mathbbm{c } \ : , \label{eq : bounddisp}\end{aligned}\ ] ] where @xmath70 $ ] . * proof * the proof is straightforward from proposition [ p : bound2mix ] with the gaussian map @xmath71 .
we also use the property of the wigner function @xmath72(\beta ) = w[d^\dag(\beta)\varrho d(\beta)](0 ) \:,\ ] ] and @xmath73 .
@xmath50 for any single - mode quantum state @xmath3 belonging to the gaussian convex hull @xmath27 , the following inequality holds @xmath6(0 ) \geq
\max_{\xi \in \mathbbm c } \left ( \frac2\pi \exp\ { -2 \bar{n}_\xi ( 1+\bar{n}_\xi)\ } \right)\ : , \label{eq : boundsq}\end{aligned}\ ] ] where @xmath74 $ ] . * proof * the proof follows from proposition [ p : bound2mix ] , by considering the gaussian map @xmath75 .
moreover , since the value of the wigner function at the origin is invariant under any squeezing operation , _ i.e. _ @xmath76(0 ) = w[\varrho](0 ) \:,\end{aligned}\ ] ] one can maximize the rhs of inequality ( [ eq : bound2mix ] ) with regard to the squeezing parameter @xmath56 .
@xmath50 + the violation of any of the inequalities presented in the last two propositions and two corollaries provides a _ sufficient _ condition to conclude that a state is quantum non - gaussian .
we formalize this by re - expressing the previous results in the form of two criteria for the detection of quantum non - gaussianity .
[ c : uno ] let us consider a quantum state @xmath3 and define the quantity @xmath77 = w[\varrho](0 ) - \frac2\pi \exp\{-2 \bar{n}(\bar{n}+1)\ } \ : .
\label{eq : delta1}\end{aligned}\ ] ] then , @xmath77 < 0 \:\ :
\rightarrow\:\ : \varrho \notin \mathcal{g } , \nonumber \end{aligned}\ ] ] that is , @xmath3 is quantum non - gaussian .
[ c : due ] let us consider a quantum state @xmath3 , a gaussian map @xmath65 ( or a convex mixture thereof ) , and define the quantity @xmath78 = w[\mathcal{e}_{\sf g}(\varrho)](0 ) - \frac2\pi \exp\{-2 \bar{n}_\mathcal{e}(\bar{n}_\mathcal{e}+1)\}\ : .
\label{eq : delta2}\end{aligned}\ ] ] then , @xmath79 < 0 \rightarrow \ : \varrho \notin \mathcal{g}. \nonumber \ ] ] typically , criterion [ c : uno ] can be useful to detect quantum non - gaussianity of phase - invariant states having the minimum of the wigner function at the origin of phase space . on the other hand , criterion [ c : due ] is of broader applicability . to give two paradigmatic examples
, the latter criterion can be useful if : ( i ) the minimum of the wigner function is far from the origin , so that one may be able to violate inequality ( [ eq : bounddisp ] ) by considering displacement operations ; ( ii ) the state is not phase - invariant and presents some squeezing , and thus one may be able to violate inequality ( [ eq : boundsq ] ) by using single - mode squeezing operations .
in this section we test the effectiveness of our criteria , by applying them to typical quantum states that are of relevance to the quantum optics community .
we shall consider pure , non - gaussian states evolving in a lossy channel , and test their quantum non - gaussianity after such evolution .
specifically , we focus on the family of quantum channels associated to the markovian master equation @xmath80 the resulting time evolution , characterized by the parameter @xmath81 , models both the incoherent loss of photons in a dissipative zero temperature environment , and inefficient detectors with an efficiency parameter @xmath82
. the evolved state @xmath83 can be equivalently derived by considering the action of a beam splitter with reflectivity @xmath84 , which couples the system to an ancillary mode prepared in a vacuum state .
the corresponding average photon number reads @xmath85 = ( 1-\epsilon)\ : \bar{n}_0 \:,\end{aligned}\ ] ] where @xmath86 $ ] is the initial average photon number .
+ it is well known that , for @xmath87 ( _ i.e. _ for detector efficiencies @xmath88 ) , no negativity of the wigner function can be observed .
we will focus then on the violation of our criteria for larger values of @xmath84 , which ensures that the evolved states have a positive wigner function .
+ notice that the quantum map @xmath89 is a gaussian map . as a consequence , by combining the divisibility property of the map ( inherited from the markovian structure of eq . ) and criterion [ c : due ] , if a violation is observed for a given loss parameter @xmath90 , then the state is quantum non - gaussian for any lower value @xmath91 @xcite .
for this reason we will focus on the maximum values of the loss parameter @xmath84 for which a violation of the bounds is observed , _
i.e. _ @xmath92 & = \max\ { \epsilon \ : : \delta_1[{\mathcal{e}_\epsilon}(\varrho ) ] \leq 0 \ } \ : , \\ { \epsilon_{\sf max}^{(2)}}[\varrho ] & = \max\ { \epsilon \ : : \exists\mathcal{e}_{\sf g}\text { s.t . } \ \delta_2[{\mathcal{e}_\epsilon}(\varrho),\mathcal{e}_{\sf g } ] \leq 0 \ } \:.\end{aligned}\ ] ] + in what follows , we start by focusing on criterion [ c : uno ] , and thus we will look for negative values of the non - gaussianity indicator @xmath93 $ ] defined in eq .
( [ eq : delta1 ] ) .
we will consider different families of states , namely fock states , photon - added coherent states and photon - subtracted squeezed states . in section [ s : violation2 ]
, we will study how to improve the results obtained , by considering the second criterion and thus by studying the non - gaussianity indicator @xmath94 $ ] .
let us start by considering fock states @xmath95 , that is the eigenstates of the number operator : @xmath96 .
a fock state evolved in a lossy channel can be written as a mixture of fock states as @xmath97 with @xmath98 we recall here that the wigner function at the origin is proportional to the expectation value of the parity operator @xmath99 , that is @xmath6 ( 0 ) = \frac2\pi { { \rm tr}}[\varrho \pi ] = \frac2\pi \left(p_{\sf even } - p_{\sf odd } \right)\ : , \label{eq : evenodd}\end{aligned}\ ] ] where @xmath100 ( @xmath101 ) represents the probability of detecting an even ( odd ) number of photons . by using eq .
( [ eq : fockev ] ) one obtains @xmath102(0 ) = \frac2\pi ( 2\epsilon -1)^m \:,\end{aligned}\ ] ] and thus the non - gaussianity indicator reads @xmath103 = \frac2\pi\left\ { ( 2\epsilon -1)^m - e^{-2(1-\epsilon ) m [ ( 1-\epsilon ) m + 1 ] } \right\ } \:.\end{aligned}\ ] ] $ ] for the first three fock states : red - dotted line : @xmath104 ; green - dashed line : @xmath105 ; blue - solid line : @xmath106 .
+ ( right ) maximum value of the noise parameter @xmath107 such that the bound ( [ eq : boundmix ] ) is violated for the state @xmath108 , as a function of the fock number @xmath109 .
[ f : fock1],title="fig : " ] $ ] for the first three fock states : red - dotted line : @xmath104 ; green - dashed line : @xmath105 ; blue - solid line : @xmath106 .
+ ( right ) maximum value of the noise parameter @xmath107 such that the bound ( [ eq : boundmix ] ) is violated for the state @xmath108 , as a function of the fock number @xmath109 .
[ f : fock1],title="fig : " ] the behavior of @xmath110 $ ] as a function of @xmath84 for the first three fock states is plotted in fig .
[ f : fock1 ] ( left panel ) .
one can observe that the criterion works really well for the fock state @xmath111 , which is proven to be quantum non - gaussian for all values of @xmath112 .
for the fock states @xmath113 and @xmath114 , a non monotonous behavior of @xmath115 is observed as a function of the loss parameter . still , negative values of the non - gaussian indicator are observed in the region of interest @xmath116 .
however , the maximum value of the noise parameter @xmath117 decreases monotonically as a function of @xmath109 , as shown in fig .
[ f : fock1 ] ( right panel ) . by increasing the fock number @xmath109
, it settles to the asymptotic value @xmath118 . as one would expect by looking at the bound in eq .
( [ eq : boundmix ] ) , for high values of the average photon number , the criterion becomes practically equivalent to the detection of negativity of the wigner function , and thus the maximum noise corresponds to @xmath119 .
a _ photon - added coherent _ ( pac ) state is defined as @xmath120 the operation of photon - addition has been implemented in different contexts @xcite , and in particular non - gaussianity and non - classicality of pac states have been investigated in @xcite .
+ being the non - gaussianity indicator @xmath93 $ ] phase insensitive , we can consider @xmath121 without loss of generality .
the average photon number can be easily calculated obtaining @xmath122 while its wigner function reads @xmath123 ( \lambda)&=\frac{2}{\pi}\frac{e^{-2(\alpha-\lambda)(\alpha-\lambda^*)}}{1+\alpha^2}\times\\ & \left(-1+\alpha^2 + 4|\lambda|^2 - 2\alpha(\lambda+\lambda^*)\right)\;.\end{aligned}\ ] ] the wigner function of the state after the loss channel @xmath124 can be evaluated by means of the formula @xmath125(\lambda)=\int \!\!d^2 \lambda ' k_\epsilon(\lambda , \lambda')w[\varrho](\lambda^\prime)\ ; , \label{eq : wigev}\end{aligned}\ ] ] where @xmath126 $ ] for pac states as a function of @xmath84 and for different values of @xmath57 : red - dotted line : @xmath127 ; green - dashed line : @xmath128 ; blue - solid line : @xmath129 .
+ ( right ) maximum value of the noise parameter @xmath107 such that the bound ( [ eq : boundmix ] ) is violated for the state @xmath124 , as a function of the parameter @xmath57 .
[ f : pacs1],title="fig : " ] $ ] for pac states as a function of @xmath84 and for different values of @xmath57 : red - dotted line : @xmath127 ; green - dashed line : @xmath128 ; blue - solid line : @xmath129 .
+ ( right ) maximum value of the noise parameter @xmath107 such that the bound ( [ eq : boundmix ] ) is violated for the state @xmath124 , as a function of the parameter @xmath57 .
[ f : pacs1],title="fig : " ] the non - gaussianity indicator @xmath130 can then be straightforwardly evaluated and is plotted in fig .
[ f : pacs1 ] ( left ) as a function of @xmath84 for different values of @xmath57 .
we note that negative values of the indicator can be observed in an interval for the noise parameter @xmath84 , which decreases with the increase of @xmath57 .
we can explain this feature by noting that , as @xmath57 decreases , the pac state approaches the fock state @xmath131 : as a consequence its quantum non - gaussianity can be more easily detected via criterion [ c : uno ] , in particular due to the minimum value of the wigner function approaching the origin of the phase space .
we plotted in fig .
[ f : pacs1 ] ( right ) the maximum value @xmath117 at which the violation of the bound is observed as a function of @xmath57 .
similarly to fock states , we observe that by increasing the energy this value tends to the asymptotic value @xmath118 .
let us consider now another important class of non - gaussian states that can be engineered with current technology .
the _ photon - subtracted squeezed _ ( pss ) states are defined as @xmath132 for low values of squeezing , these states approximate the _ schrdinger kitten _ states , that is , superpositions of coherent states @xmath133 with opposite phase and small amplitude ( @xmath134 ) @xcite .
the generation of this kind of states has been demonstrated experimentally @xcite , and it relies on performing conditional photon number measurements .
+ without loss of generality we shall consider a real squeezing parameter @xmath135 ; the corresponding average photon number of a pss state reads @xmath136 while its wigner function is @xmath137(\lambda ) & = - \frac2\pi e^{-2|\lambda|^2 \cosh{2 r } + ( \lambda^2 + \lambda^{*2 } ) \sinh{2 r } } \times \nonumber \\ & \left[1 - 4|\lambda|^2 \cosh{2 r } + 2 ( \lambda^2 + \lambda^{*2 } ) \sinh{2 r } \right ] \:.\end{aligned}\ ] ] as for the pac states , the wigner function of the evolved state can be evaluated by means of eq .
( [ eq : wigev ] ) and the non - gaussianity indicator @xmath138 $ ] can be evaluated accordingly .
its behavior as a function of @xmath84 and for different values of the squeezing factor @xmath139 is plotted in fig .
[ f : psss1 ] ( left ) .
$ ] for pss states as a function of @xmath84 and for different values of @xmath139 : red - dotted line : @xmath140 ; green - dashed line : @xmath141 ; blue - solid line : @xmath142 .
+ ( right ) maximum value of the noise parameter @xmath107 such that the bound ( [ eq : boundmix ] ) is violated for the state @xmath143 , as a function of the initial squeezing parameter @xmath139 .
[ f : psss1],title="fig : " ] $ ] for pss states as a function of @xmath84 and for different values of @xmath139 : red - dotted line : @xmath140 ; green - dashed line : @xmath141 ; blue - solid line : @xmath142 . + ( right ) maximum value of the noise parameter @xmath107 such that the bound ( [ eq : boundmix ] ) is violated for the state @xmath143 , as a function of the initial squeezing parameter @xmath139 .
[ f : psss1],title="fig : " ] in the right panel of fig .
[ f : psss1 ] we plot the maximum noise parameter @xmath117 as a function of the squeezing parameter @xmath139 , observing the same behavior obtained for fock and pac states : the value of @xmath117 decreases monotonically with the energy of the state , approaching the asymptotic value @xmath118
. we will now show how the second criterion , which is based on the violation of the inequality ( [ eq : bound2mix ] ) , can be exploited in order to improve the results shown in the previous section . since in this case one can optimize the procedure over an additional gaussian channel , in general one has @xmath144 .
the simplest gaussian maps that one can consider are displacement and squeezing operations ; correspondingly we are going to seek violation of the bounds described by eqs .
( [ eq : bounddisp ] ) and ( [ eq : boundsq ] ) . as anticipated in sec .
[ s : criteria ] , these new criteria are useful for states which are not phase invariant : the paradigmatic examples are states _ displaced _ in the phase - space , that is , having the minimum of the wigner function outside the origin , or states that exhibit squeezing in a certain quadrature . due to this fact , the bounds based on eqs .
( [ eq : bounddisp ] ) and ( [ eq : boundsq ] ) can not help in optimizing the results we obtained for fock states .
we will focus then on the other classes of states we introduced , that is pac and pss states . for @xmath145 .
the minimum of the wigner function is not at the origin of the phase - space , and the state has non - zero first moments .
[ f : wigpac0 ] ] by looking at the pac state wigner function in fig .
[ f : wigpac0 ] , one observes that its minimum is not at the origin of the phase space
. moreover , these states have non - zero first moments , implying that one can decrease their average photon number by applying an appropriate displacement .
both observations suggest that it is possible to decrease the value the quantum non - gaussianity indicator defined in eq .
( [ eq : delta2 ] ) , @xmath146 \ : , \label{eq : deltapac}\end{aligned}\ ] ] by means of a displacement operation @xmath147 . to evaluate @xmath148 according to eq .
( [ eq : deltapac ] ) one has simply to evaluate the wigner function of the state @xmath149 in a displaced point in the phase space , _ i.e. _ @xmath150(-\beta)$ ] , and its average photon number @xmath151 where @xmath152 , and for @xmath153 , @xmath154 our goal is then to minimize @xmath148 over the possible displacement parameters @xmath155 .
+ as a function of the additional displacement parameter @xmath155 , for @xmath156 and for different values of the initial parameter @xmath57 : : red - dotted line : @xmath127 ; green - dashed line : @xmath128 ; blue - solid line : @xmath129 .
+ ( right ) optimized non - gaussianity indicator @xmath157 as a function of @xmath84 and for different values of @xmath57 , where the displacement parameter @xmath158 has been chosen as in eq . ( [ eq : betaopt ] ) : @xmath127 ; green - dashed line : @xmath128 ; blue - solid line : @xmath129 .
[ f : pacs2],title="fig : " ] as a function of the additional displacement parameter @xmath155 , for @xmath156 and for different values of the initial parameter @xmath57 : : red - dotted line : @xmath127 ; green - dashed line : @xmath128 ; blue - solid line : @xmath129 .
+ ( right ) optimized non - gaussianity indicator @xmath157 as a function of @xmath84 and for different values of @xmath57 , where the displacement parameter @xmath158 has been chosen as in eq . ( [ eq : betaopt ] ) : @xmath127 ; green - dashed line : @xmath128 ; blue - solid line : @xmath129 .
[ f : pacs2],title="fig : " ] in fig .
[ f : pacs2 ] ( left ) we plot @xmath148 as a function of @xmath155 for different values of the coherent state parameter @xmath57 and for @xmath156 . we observe that , while for @xmath159 the bound is not always violated , it is possible to find values such that @xmath160 and thus prove that the state is quantum non - gaussian .
unfortunately the optimal value @xmath158 , which minimizes @xmath148 , can not be obtained analytically .
however we observed that for large values of @xmath84 and for @xmath161 one can approximate it as @xmath162 the behavior of @xmath157 as a function of @xmath84 shown in fig .
[ f : pacs2 ] ( right ) , for different values of @xmath57 and fixing @xmath158 as in eq . .
if we compare this with fig .
[ f : pacs1 ] , not only we observe an improvement in our capacity to witness quantum non - gaussianity for these states , but we also see that @xmath157 remains negative for all values of @xmath84 . indeed , numerical investigations seem to suggest that @xmath163 for all the possible values of @xmath57 : we indeed conjecture that any initial pac state remains quantum non - gaussian during the lossy evolution induced by eq . , and that this feature can be captured by our second criterion .
however , as one can observe from fig .
[ f : pacs2 ] ( right ) , the non - gaussianity indicator approaches zero quite fast with both @xmath57 and @xmath84 , and thus it may be more challenging to detect its negativity in an actual experiment for states with a high average photon number and for large losses . for @xmath141 .
the minimum of the wigner function is at the origin of the phase - space , and the state exhibits squeezing in one of the quadratures .
[ f : wigpss0 ] ] like pac states inherit a displacement in phase space from the initial coherent states , pss states inherit squeezing , as we can observe by looking at the wigner function in fig . [
f : wigpss0 ] .
this motivates us to make use of corollary [ c : due ] , and thus optimize the non - gaussianity indicator in eq .
( [ eq : delta2 ] ) as @xmath164 \ : , \label{eq : deltapss}\end{aligned}\ ] ] that is by considering an additional squeezing operation @xmath165 on the evolved state @xmath166 . as pointed out in the proof of inequality ( [ eq : boundsq ] ) , the wigner function at the origin is invariant under squeezing operations .
hence , the optimal value @xmath167 that minimizes @xmath168 coincides with the value which minimizes the average photon number of @xmath169 , @xmath170 + \nu_s^2\,,\end{aligned}\ ] ] where @xmath171 , @xmath172 and for an initial pss state ( with a real squeezing parameter @xmath139 ) , @xmath173 as a function of the additional squeezing parameter @xmath174 , for @xmath175 and for different values of the initial parameter @xmath139 : red - dotted line : @xmath140 ; green - dashed line : @xmath141 ; blue - solid line : @xmath142 .
+ ( right ) optimized non - gaussianity indicator @xmath176 as a function of @xmath84 and for different values of @xmath139 , where the squeezing parameter is given by : red - dotted line : @xmath140 ; green - dashed line : @xmath141 ; blue - solid line : @xmath142 .
[ f : psss2],title="fig : " ] as a function of the additional squeezing parameter @xmath174 , for @xmath175 and for different values of the initial parameter @xmath139 : red - dotted line : @xmath140 ; green - dashed line : @xmath141 ; blue - solid line : @xmath142 .
+ ( right ) optimized non - gaussianity indicator @xmath176 as a function of @xmath84 and for different values of @xmath139 , where the squeezing parameter is given by : red - dotted line : @xmath140 ; green - dashed line : @xmath141 ; blue - solid line : @xmath142 .
[ f : psss2],title="fig : " ] the behavior of @xmath168 as a function of the additional squeezing @xmath174 is plotted in fig .
[ f : psss2 ] . as we observed in the previous case , the optimised criterion works in cases where the bound ( [ eq : boundmix ] ) ( corresponding to @xmath177 ) was not violated .
+ moreover the optimal squeezing value can be evaluated analytically , yielding @xmath178 and @xmath107 , obtained respectively by means of the optimized and not - optimized criteria , for the state @xmath143 , as a function of the initial squeezing parameter @xmath139 .
red - dotted line : @xmath179 ; blue - solid line : @xmath117 .
[ f : psss3 ] ] the optimized quantum non - gaussianity indicator @xmath176 is plotted in fig .
[ f : psss2 ] ( right ) , where we observe that negative values are obtained for large values of losses . however , while for pac states we had evidence that the maximum value of losses is @xmath163 for all the possible initial states , this is no longer true for pss states . the behavior of @xmath179 as a function of @xmath139 is plotted in fig .
[ f : psss3 ] , together with the previously obtained @xmath117 .
we can notice the big improvement in our detection capability , obtained by exploiting corollary [ c : due ] ; however for large values of @xmath139 we still observe that @xmath179 decreases towards the same limiting value @xmath180 .
moreover , as it can be observed in fig .
[ f : psss2 ] ( right ) , the indicator @xmath181 approaches zero by increasing @xmath139 and @xmath84 , and thus also in this case it can become challenging to witness quantum non - gaussianity with our methods , in experiments with large values of the initial squeezing @xmath139 and large losses .
we have presented a set of criteria to detect quantum non - gaussian states , that is , states that can not be expressed as mixtures of gaussian states .
the first criterion is based on seeking the violation of a lower bound for the values that the wigner function can take at the origin , depending only on the average photon number of the state . to verify the effectiveness of the criterion , we considered the evolution of non - gaussian pure states in a lossy gaussian channel , looking for the maximum value of the noise where such bound is violated .
we observed that the criterion works well , detecting quantum non - gaussianity in the non - trivial region of the noise parameters where no negativity of the wigner function can be observed .
we have also shown how the criterion can be generalized and improved , by optimising over additional gaussian operations applied to the states of interest .
notice that in a possible experimental implementation one does not need to perform such additional gaussian operations , such as displacement or squeezing , in the actual experiment .
indeed , it suffices to use the data obtained on the state itself , and then apply suitable post - processing to evaluate the optimized non - gaussianity indicator .
our criterion , which expresses a sufficient condition for quantum non - gaussianity , shares some similarities with hudson s theorem for pure gaussian states , in the sense that it establishes a relationship between the concept of gaussianity ( combined with classical mixing ) , and the possible values that a wigner function can take .
the successful implementation of our criteria corresponds to the measurement of the wigner function at the origin of the phase space which , in turn , corresponds to the ( photon ) parity of the state under investigation .
this may be obtained with current technology by direct parity measurement @xcite , or by reconstruction of the photon distribution either by tomographic reconstruction or by the on / off method @xcite . when the criterion is satisfied
, one can confirm that the quantum state at disposal has been generated by means of a highly non - linear process , even in the cases where , perhaps due to inefficient detectors or other types of noise , negativity of the wigner function can not be detected .
mgg , tt and msk thank radim filip for discussions .
mgap thanks vittorio giovannetti for discussions .
mgg acknowledges support from uk epsrc ( ep / i026436/1 ) .
t.t . and m.s.k .
acknowledge support from the nprp 4- 426 554 - 1 - 084 from qatar national research fund .
so and mgap acknowledge support from miur ( firb `` lichis '' no .
rbfr10yq3h ) .
99 e. p. wigner , phys . rev . * 40 * , 749 ( 1932 ) r. glauber , phys . rev . * 131 * , 2766 ( 1963 ) .
e. c. g. sudarshan , phys .
* 10 * , 277 ( 1963 ) . c. t. lee , phys .
a * 44 * , r2775 ( 1991 ) . c. t. lee , phys .
a * 45 * , 6586 ( 1992 ) .
arvind , n. mukunda , r. simon , phys.rev .
a * 56 * , 5042 ( 1997 ) .
arvind , n. mukunda , r. simon , j. phys a * 31 * : 565 ( 1998 ) .
m. a. marchiolli , v. s. bagnato , y. guimaraes , b. baseia , phys .
a * 279 * , 294 ( 2001 ) .
m. g. a. paris , phys .
a * 289 * , 167 ( 2001 ) .
dodonov , j. opt .
b * 4 * , r1 , ( 2002 ) .
a. kenfack , k. zyczkowski , j. opt .
b * 6 * , 396 ( 2004 )
. e. v. shchukin , w. vogel , phys . rev . a * 72
* , 043808 ( 2005 ) .
t. kiesel , w. vogel , v. parigi , a. zavatta , m. bellini , phys . rev . a * 78 * , 021804 r ( 2008 ) . w. vogel , phys .
lett . * 100 * , 013605 ( 2008 ) .
r. simon , phys .
. lett . * 84 * , 2726 ( 2000 ) .
lu - ming duan , g. giedke , j.i .
cirac , p. zoller , phys .
* 84 * , 2722 ( 2000 ) .
p. marian , t. a. marian , h. scutaru , phys .
lett . * 88 * , 153601 ( 2002 ) .
p. giorda , m. g. a. paris , phys .
lett . * 105 * , 020503 ( 2010 ) ; g. adesso , a. datta phys . rev . lett . * 105 * , 030501 ( 2010 ) .
a. ferraro , m. g. a. paris , phys .
lett * 108 * , 260403 ( 2012 ) . c. gehrke , j. sperling , w. vogel , phys .
a * 86 * , 052118 ( 2012 ) .
d. buono , g. nocerino , v. dauria , a. porzio , s. olivares , m.g.a .
paris , j. opt .
b * 27 * , 110 ( 2010 ) ; d. buono , g. nocerino , a. porzio , s. solimeno .
phys . rev .
a * 86 * , 042308 ( 2012 ) .
m. g. genoni , m. g. a. paris and k. banaszek , phys .
a * 76 * , 042327 ( 2007 ) .
m. g. genoni , m. g. a. paris and k. banaszek , phys .
a * 78 * , 060303 ( 2008 ) .
m. g. genoni and m. g. a. paris , phys .
a * 82 * , 052341 ( 2010 ) .
m. barbieri , n. spagnolo , m. g. genoni , f. ferreyrol , r. blandino , m. g. a. paris , p. grangier and r. tualle - brouri , phys .
a * 82 * , 063833 ( 2010 ) .
r. filip , l. mista , jr .
lett . * 106 * , 200401 ( 2011 ) .
m. jeek , i. straka , m. miuda , m. duek , j. fiuraek , r. filip , phys .
107 * , 213602 ( 2011 ) .
m. jeek , a. tipsmark , r. dong , j. fiuraek , l. mita , jr .
, r. filip , u. l. andersen , phys . rev .
a * 86 * , 043813 ( 2012 ) .
a. predojevic , m. jezek , t. huber , h. jayakumar , t. kauten , g. s. solomon , r. filip and g. weihs , arxiv:1211.2993 [ quant - ph ] .
r. filip , phys .
a * 87 * , 042308 ( 2011 ) . v. dauria , c. de lisio , a. porzio , s. solimeno , j. anwar , m. g. a. paris , phys . rev . a * 81 * , 033846 ( 2010 ) .
r. l. hudson , rep .
* 6 * , 249 ( 1974 ) .
f. soto and p. claverie , j. math .
* 24 * , 97 ( 1983 ) .
a. mandilara , e. karpov and n. j. cerf , phys .
a * 79 * , 062302 ( 2009 ) .
u. m. titulaer and r. j. glauber , phys . rev . *
140 * , 676 ( 1963 ) .
a. mari and j. eisert , phys
lett . * 109 * , 230503 ( 2012 ) .
v. veitch , n. wiebe , c. ferrie and j. emerson , new j. phys .
* 15 * , 013037 ( 2013 ) .
j. heersink , c. marquardt , r. dong , r. filip , s. lorenz , g. leuchs and u. l. andersen , phys .
lett . * 96 * , 253601 ( 2006 ) .
k. e. cahill and r. j. glauber , phys . rev .
* 177 * , 1882 ( 1969 ) .
j. eisert and m. m. wolf , _ quantum information with continuous variables of atoms and light _ , edited by n. j. cert , g. leuchs and e. s. polzik ( imperial college press , london , 2007 ) , pp .
being the map @xmath89 divisible , then for all @xmath91 , a parameter @xmath182 exists , such that @xmath183 . as a consequence ,
if a criterion is violated for the quantum state @xmath184 , the quantum state @xmath185 is quantum non - gaussian for all @xmath186 .
a. zavatta , s. viciani and m. bellini , phys .
a * 70 * , 053821 ( 2004 ) .
a. zavatta , v. parigi and m. bellini , phys . rev .
a * 75 * , 052106 ( 2007 ) .
v. parigi , a. zavatta , m. s. kim and m. bellini , science * 317 * , 1890 ( 2007 ) .
a. zavatta , v. parigi , m. s. kim , h. jeong and m. bellini , phys .
* 103 * , 140406 ( 2009 ) .
m. dakna , t. anhut , t. opatrny , l. knoll and d .-
welsch , phys .
a * 55 * , 3184 ( 1997 ) ; m. s. kim , e. park , p. l. knight and h. jeong , phys
a * 71 * , 043805 ( 2005 ) ; s. olivares and m. g. a. paris , j. opt b * 7 * , 616 ( 2005 ) .
a. ourjoumtsev , r. tualle - brouri , j. laurat and p. grangier , science * 312 * , 83 ( 2006 ) .
k. wakui , h. takahashi , a. furusa and m. sasaki , opt .
express * 15 * , 3568 ( 2007 ) .
j. s. neergaard - nielsen , b. m. nielsen , c. hettich , k. molmer and e. s. polzik , phys .
97 * , 083604 ( 2006 ) .
t. gerrits , s. glancy , t. s. clement , b. calkins , a. e. lita , a. j. miller , a. l. midgall , s. w. nam , r. p. mirin and e. knill , phys .
a * 82 * , 031802 ( 2010 ) .
s. haroche , m. brune and j. m. raimond , j. mod . opt . * 54 * , 2101 ( 2007 ) .
s. wallentowitz , w. vogel , phys .
a * 53 * , 4528 ( 1996 ) .
k. banaszek , k. wodkiewicz , phys .
* 76 * , 4344 ( 1996 ) .
t. opatrny and d. g. welsch , phys .
a * 55 * , 1462 ( 1997 ) ; t. opatrny , d. g. welsch , and w. vogel , phys . rev .
a * 56 * , 1788 ( 1997 ) .
m. munroe , d. boggavarapu , m.e .
anderson , and m. g. raymer , phys .
a * 52 * , r924 ( 1995 ) ; y. zhang , k. kasai , and m. watanabe , opt . lett . *
27 * , 1244 ( 2002 ) .
m. raymer and m. beck , in _ quantum states estimation _ , lect . not .
* 649 * ( springer , berlin - heidelberg , 2004 ) .
d. mogilevtsev , opt . commun . * 156 * , 307 ( 1998 ) ; acta phys .
slovaca * 49 * , 743 ( 1999 ) .
a. r. rossi , s. olivares , m. g.a .
paris , phys .
a * 70 * , 055801 ( 2004 ) ; a. r. rossi , m. g. a. paris , eur .
. j. d * 32 * , 223 ( 2005 ) .
g. zambra , a. andreoni , m. bondani , m. gramegna , m. genovese , g. brida , a. rossi and m. g. a. paris , phys .
* 95 * , 063602 ( 2005 ) ; g. zambra , m. g. a. paris , phys . rev .
a * 74 * , 063830 ( 2006 ) .
a. allevi , a. andreoni , m. bondani , g. brida , m. genovese , m. gramegna , p. traina , s. olivares , m. g. a. paris , g. zambra , phys .
a * 80 * , 022114 ( 2009 ) .
l. a. jiang , e. a. dauler , j. t. chang , phys .
rev . a * 75 * 062325 ( 2007 ) .
a. divochiy et al . , nat
* 2 * , 302 ( 2008 ) .
d. achilles et al .
, opt . lett . * 28 * , 2387 ( 2003 ) , m. j. fitch , b. c. jacobs , t. b. pittman , j. d. franson , phys . rev .
a * 68 * , 043814 ( 2003 ) .
g. zambra , m. bondani , a. s. spinelli , a. andreoni , rev .
instrum . * 75 * , 2762 ( 2004 ) .
m. ramilli , a. allevi , a. chmill , m. bondani , m. caccia , a. andreoni , j. opt .
b , * 27 * , 852 ( 2010 ) .
j. kim , s. takeuchi , y. yamamoto , h.h .
hogue , appl .
lett . * 74 * , 902 ( 1999 ) . | we introduce a family of criteria to detect quantum non - gaussian states of a harmonic oscillator , that is , quantum states that can not be expressed as a convex mixture of gaussian states .
in particular we prove that , for convex mixtures of gaussian states , the value of the wigner function at the origin of phase space is bounded from below by a non - zero positive quantity , which is a function only of the average number of excitations ( photons ) of the state . as a consequence ,
if this bound is violated then the quantum state must be quantum non - gaussian .
we show that this criterion can be further generalized by considering additional gaussian operations on the state under examination .
we then apply these criteria to various non - gaussian states evolving in a noisy gaussian channel , proving that the bounds are violated for high values of losses , and thus also for states characterized by a positive wigner function . |
one of the current topics in solid - state magnetism is multiple - spin ordering without any single - spin dipole moment.@xcite vector and scalar chiral orders and spin multipolar orders are typical examples expected to appear in real compounds .
the emergence of vector and scalar chiralities accompanies the spontaneous breakdown of parity or time - reversal symmetries and these two could be detected indirectly from parity- or time - reversal - odd observables and related quantities ( e.g. , electric polarization in multiferroics , asymmetric momentum dependence of spin structure factors , hall conductivity , etc ) . on the other hand ,
spin multipolar orders , for example , spin nematic and spin triatic orders , in spin-@xmath1 magnets can occur without breaking spatial symmetry , since they are characterized as condensation of bound multi - magnons@xcite or spin - triplet resonating valence bond state.@xcite the absence of both spin long - range order and lattice symmetry breaking makes it difficult to find them in experiments .
furthermore , no clear experimental proof of the spin multipolar orders has ever been reported , because there is no established experimental method of probing spin multipolar orders .
thanks to recent theoretical studies , it has been gradually recognized that magnetic multipolar states occur in several geometrically frustrated spin-@xmath1 magnets , especially , in low dimensions .
@xcite in this paper , we focus on spin-@xmath1 frustrated chains with ferromagnetic ( fm ) nearest - neighboring exchange @xmath5 and competing antiferromagnetic ( af ) next - nearest - neighboring exchange @xmath6 . the hamiltonian is given by @xmath7 where @xmath8 is the spin-@xmath1 operator on the @xmath9th site , and @xmath10 is the applied magnetic field along the @xmath11 axis .
recent studies revealed that this model possesses a series of field - induced multipolar tomonaga - luttinger ( tl ) liquid phases in high magnetization regime.@xcite these multipolar tl liquids are interpreted as a hard - core bose gas of multi - magnon bound states,@xcite where the number @xmath12 of magnons forming a multi - magnon bound state changes successively from @xmath13 to 4 with varying @xmath14 . in the multipolar phase with @xmath13 , the quadrupolar ( or spin nematic ) operator @xmath15 exhibits a quasi long - range order .
similarly , the octupolar ( or triatic ) operator @xmath16 and the hexadecapolar operator @xmath17 show a quasi long - range order , respectively , in the phases with @xmath18 and @xmath19 . in all of these phases
, the longitudinal spin correlation also decays algebraically , while the transverse spin correlation is short - ranged .
these characteristic properties of the multipolar tl liquid phases are distinct from those in the standard tl liquid ( e.g. , spin-@xmath1 heisenberg chain ) .
numerical studies@xcite of the quadrupolar and octupolar tl liquid phases showed that the relevant multipolar correlations are dominant in the high - magnetization regime , whereas the longitudinal spin - density - wave ( sdw ) correlation becomes dominant in the lower - magnetization regime .
we call the high - field multipolar states with @xmath13 and 3 _ quadrupolar and octupolar liquids _ and the low - field multipolar states _ sdw@xmath20 and sdw@xmath21 liquids _ , respectively , following ref . . in these multipolar tl liquids ,
bulk static quantities , such as entropy , specific heat , and uniform susceptibility , exhibit qualitatively the same behaviors as those of standard tl liquids , and hence it is impossible to distinguish multipolar and standard tl liquids from them . to detect direct evidence of their ordering , we need to measure multiple - spin order parameter or relevant multiple - spin correlation functions , which is generally difficult . instead , to find any indirect evidence , it is necessary to specify an effective method of detecting any signature of spin multipolar ordering through usual experiments .
recently , ref . has proposed an experimental way of detecting a signature of multipolar tl liquids , showing that the nuclear magnetic resonance ( nmr ) relaxation rate @xmath0 decreases with lowering temperature @xmath22 in the multipolar tl liquids near saturation .
this is completely different from the behavior of standard tl liquids in one - dimensional ( 1d ) af magnets such as the spin-@xmath1 heisenberg chain , in which @xmath0 always increases with lowering @xmath22 irrespective of the value of applied magnetic field @xmath10 .
however , it is generally hard to experimentally approach the vicinity of the saturation field , especially , in magnets with strong exchange couplings .
in fact , the saturation field of a @xmath2-@xmath3 magnet @xmath4 is about @xmath23 t , @xcite and to perform nmr measurement under such a high field is not an easy task .
an experimentally detecting scheme at low magnetic field is therefore desirable .
it was also shown in ref . that the momentum dependence of dynamical structure factors in the multipolar tl liquids shows a distinct difference from that of the spin-@xmath1 af heisenberg chain . in this paper , we reexamine the nmr relaxation rate @xmath0 in the spin quadrupolar ( nematic ) and sdw@xmath20 tl liquids , including both low and high - magnetic - field regimes , and evaluate its field and temperature dependence quantitatively . to this end , we combine the field - theoretical ( bosonization ) approach with the density - matrix renormalization - group ( dmrg ) method , substituting numerical values for nonuniversal parameters in analytic results .
we compare these results with the relaxation rate @xmath0 of the spin-@xmath1 af heisenberg chain .
it is found that , in addition to the temperature dependence , the field dependence of @xmath0 in the nematic and sdw@xmath20 liquids clearly differs from that in the usual tl liquid of the spin-@xmath1 af chain . in the @xmath2-@xmath3 chain ,
the relaxation rate @xmath0 slowly decreases with increase of magnetic field ( or is almost independent of the field ) in the low - field sdw@xmath20 regime and increases suddenly in the high - field nematic regime near saturation , whereas it increases monotonically , as a function of the field , in the usual tl liquid of the spin-@xmath1 af chain .
the non - monotonic field dependence of @xmath0 is a unique characteristic of the multipolar tl liquids in spin-@xmath1 @xmath2-@xmath3 chains .
furthermore , if we tune the direction of the field @xmath10 , we can eliminate the dominant contribution to @xmath0 in the multipolar phases and thereby make @xmath0 show exponential decays at low temperature in both nematic and sdw@xmath20 liquids .
this exponential decay is caused by a spin gap in the spin transverse excitations , which is also a unique property of the multipolar tl liquids .
our prediction would be applicable in nmr experiments of quasi one - dimensional ( 1d ) edge - sharing cuprate magnets , such as @xmath4 ( refs . ) , @xmath24 ( ref . ) , @xmath25 ( refs . ) , @xmath26 ( refs . ) and @xmath27 ( ref . ) , whose magnetic properties are expected to be described by the spin-@xmath28 @xmath2-@xmath3 model with fm @xmath2 and af @xmath3 .
the paper is organized as follows . in secs .
[ sec : nmr ] and [ sec : multipolar ] , we review briefly the theory of the nmr relaxation rate @xmath0 in electron spin systems and effective theories for multipolar and usual tl liquids . particularly , in sec .
[ sec : multipolar ] , relying on an established field - theoretical method , we write down the formula of @xmath0 in multipolar liquid phases of the @xmath2-@xmath3 chain ( [ eq : zigzag ] ) and also that in tl liquid of the spin-@xmath1 af heisenberg chain .
section [ sec : h - t - dep ] contains the main results of the present paper , which are quantitative estimates of @xmath0 in both quadrupolar and usual tl liquids as a function of magnetic field and temperature .
section [ sec : realmagnets ] is devoted to a discussion of some relevant factors in real magnets which are neglected in the previous sections .
we consider a broad temperature scale beyond the low - energy effective theory , the relation between electron - nuclear spin interaction and direction of field @xmath10 , and the effects of dzyaloshinsky - moriya interaction .
tuning of the field direction is also discussed . finally , in sec .
[ sec : conclusions ] , we summarize our results and their relevance to real quasi-1d compounds .
here , we briefly explain the formula of the nmr relaxation rate @xmath0 in electron - spin systems . provided that the nmr relaxation process is mainly caused by interaction between electron spin @xmath29 and nuclear spin @xmath30 , the standard perturbation theory for the interaction evaluates the relaxation rate @xmath0 as follows : @xcite @xmath31 \nonumber\\ & & + |\tilde a_\parallel(k)|^2{\cal s}^{zz}(k,\omega).\end{aligned}\ ] ] here
, we have assumed that the electron - spin system is in one dimension .
the electron - spin dynamical structure factor at finite temperature @xmath22 is defined as @xmath32 where @xmath33 denotes the thermal average .
the frequency @xmath34 in eq .
( [ eq:1/t_1 ] ) is a given resonant value of applied oscillating field . since
its magnitude is generally much smaller than the energy scale of electron systems , we may take a limit @xmath35 in eq .
( [ eq:1/t_1 ] ) .
the symbols @xmath36 and @xmath37 denote fourier components of hyperfine coupling constants between electron and nuclear spins , which generally stem from the dipole - dipole interaction and an su(2)-invariant exchange interaction @xmath38 .
the longitudinal component @xmath39 originates only from the dipole - dipole interaction . in quantum spin systems
, @xmath39 is usually the same order as @xmath40 .
the spatial range of interactions between electron spins and a single nuclear spin is local , that is , at most the order of the lattice spacing @xmath41 .
therefore , the @xmath42 dependence of @xmath43 could be negligible . under the assumption of such a locality , the hyperfine coupling term in the hamiltonian
is given by @xmath44 where the electron site closest to the nucleus is assumed to be @xmath45 and @xmath46 is the real - space hyperfine coupling tensor . in this case
, @xmath0 can be approximated as @xmath47 \nonumber\\ & & + a^2_\parallel(j=0){\cal s}^{zz}_{j=0}(\omega).\end{aligned}\ ] ] here , hyperfine couplings @xmath48 and @xmath49 are proportional to @xmath50 and @xmath51 , respectively .
the local dynamical structure factor @xmath52 at finite temperature is represented as @xmath53
in this section , we briefly review the effective theory for multipolar tl - liquid phases in the @xmath2-@xmath3 spin chain ( [ eq : zigzag ] ) and the well - established tl - liquid theory in the spin-@xmath1 af heisenberg chain . using these theories
, we can derive analytic forms of dynamical structure factors and @xmath0 .
two analytic approaches have been developed for field - induced multipolar phases in the model ( [ eq : zigzag ] ) : one is a weak - coupling approach @xcite where @xmath54 is assumed and @xmath2 is treated as a perturbation for two decoupled af-@xmath3 chains .
the other is a phenomenological bosonization in the vicinity of the saturation field .
@xcite the former weak - coupling theory can lead to only the quadrupolar ( nematic ) phase , but it enables us to calculate various physical quantities in principle .
the latter approach can treat all of the multipolar phases ( quadrupolar , octupolar , and hexadecapolar ) , while the quantities that can be evaluated are somewhat restricted . here , we employ the latter approach , which is sufficient to calculate @xmath0 ( the former will be used in sec .
[ sec : realmagnets ] ) . in the @xmath2-@xmath3 chain with ferromagnetic @xmath55 , the excitation mode which destabilizes
the fully polarized state at the saturation field @xmath56 is a multimagnon bound state consisting of @xmath57 magnons . below the saturation field ,
the soft bound magnons proliferate to form tl liquid with dominant transverse multipolar correlation .
the system is thus described as a hard - core bose gas of @xmath12-magnon bound states . from a numerical calculation of the @xmath12-magnon excitation energies
, it has been shown that the soft mode to realize the multipolar tl liquid is two- , three- , and four - magnon bound states with momentum @xmath58 for @xmath59 , @xmath60 , and @xmath61 , respectively.@xcite from the picture of the hard - core bose gas , we may map multipolar operators and the @xmath11 component of spin operators to operators of bosons in the following forms @xmath62 where @xmath63 is an annihilation operator of the hard - core boson , and we have considered the @xmath12th - order multipolar liquid states ( @xmath64 and @xmath65 respectively correspond to quadrupolar , octupolar , and hexadecapolar liquids ) .
the staggered factor @xmath66 in the first line comes from the momentum @xmath58 of the soft multimagnon bound states . to the 1d hard - core bose gas
, we can apply the standard abelian bosonization .
@xcite its low - energy effective hamiltonian is represented as @xmath67,\end{aligned}\ ] ] where @xmath68 is a spatial coordinate , @xmath69 and @xmath70 are a canonical pair of scalar fields , @xmath71 is the tl - liquid parameter , and @xmath72 is the elementary excitation velocity .
we set @xmath73 here at the saturation limit @xmath74 where the hard - core boson gas becomes equal to a free fermion gas .
it is numerically shown @xcite that the tl - liquid parameter @xmath71 almost decreases monotonically with lowering magnetic field @xmath10 in quadrupolar and octupolar liquid phases .
the velocity @xmath72 is expected to be of the order of @xmath75 , except for the saturation limit @xmath74 where @xmath76 .
we note that the tl - liquid theory is valid only in the temperature regime of @xmath77 .
the boson operator @xmath63 and the density operator @xmath78 are expressed in terms of @xmath69 and @xmath70 : @xcite @xmath79 , \nonumber\\ n_j & \approx & \rho+\frac{a}{\sqrt{\pi}}\partial_x\phi + g_0 \cos(\sqrt{4\pi}\phi+2\pi\rho j)+\cdots,\end{aligned}\ ] ] where @xmath80 is the mean value of the boson density and @xmath81 is the magnetization per site .
factors @xmath82 and @xmath83 are nonuniversal constants . in the present representation , the scaling dimensions of @xmath84 and @xmath85 are , respectively , @xmath86 and @xmath87 . from eqs .
( [ eq : multipolar_boson])-([eq : boson_phase ] ) , two - point correlation functions of @xmath88 and @xmath12-th multipolar operator @xmath89 at zero temperature are shown to asymptotically behave like @xmath90 \nonumber\\ & + \frac{g_0 ^ 2}{2}\cos(2\pi\rho j ) \left(\frac{a}{|w|}\right)^{2\kappa}+\cdots , \nonumber\\ \langle m_j^{(p)}(\tau)m_0^{(p)\dag}(0)\rangle_{0 } & \approx ( -1)^jf_0 ^ 2\left(\frac{a}{|w|}\right)^{1/(2\kappa)}+\cdots\end{aligned}\ ] ] in the @xmath12-th multipolar tl - liquid phase , where @xmath91 denotes the expectation value of observables at zero temperature .
( incidentally , the @xmath92 case corresponds to the usual tl liquid in spin-@xmath1 af chain .
see the following subsection . ) here , @xmath93 is imaginary time , and light - cone coordinates @xmath94 and @xmath95 are defined as @xmath96 . equation ( [ eq : correaltion_multipolar ] ) has been numerically confirmed , @xcite and this asymptotic behavior survives down to the phase boundary between multipolar and lower - field vector chiral phases .
the power - law behavior also indicates that in the high - field regime ( narrowly - defined multipolar regime ) with @xmath97 , the multipolar correlation is stronger than the longitudinal spin correlation , while the latter is dominant in the lower - field sdw regime with @xmath98 .
namely , @xmath99 is the crossover point between multipolar and sdw regimes . in the multipolar tl liquid phase
, the transverse spin correlation decays exponentially , since a finite energy is necessary to violate @xmath12-magnon bound states .
though we can not compute the transverse spin correlation function and the multipolar correlation functions @xmath100 with @xmath101 within the theory of hard - core bose gas of multimagnon bound states , we can explicitly derive the form of the transverse spin correlation function @xmath102 , which is short ranged , @xcite using the weak - coupling theory in the quadrupolar - liquid phase ( @xmath13 ) .
furthermore , the exponential decay of the @xmath103 ( @xmath104 ) magnon correlations including the transverse spin correlation has been confirmed numerically for the entire regime of the quadrupolar and octupolar tl liquid phases.@xcite the single - magnon gap can be determined by measuring transverse dynamical structure factors @xcite @xmath105 , for example , in neutron scattering experiment .
it would be at most the order of @xmath106 or @xmath3 . this fact of the short - ranged transverse spin correlation in the multipolar tl liquids plays a key role in the following argument .
the slowly decaying mode in the longitudinal spin correlator in eq . ( [ eq : correaltion_multipolar ] ) contributes to the leading temperature dependence of nmr relaxation rate @xmath0 . through an established method based on conformal field theory ,
@xcite we obtain the local dynamical structure factors at finite temperature from correlation functions at zero temperature and achieve the formula for @xmath0 @xmath107 where @xmath108 is the beta function and @xmath109 .
the first and second terms on the right - hand side are , respectively , contributions from the second and third terms of @xmath110 in eq .
( [ eq : correaltion_multipolar ] ) .
if we take into account the spatial dependence of @xmath111 instead of the formula ( [ eq:1/t_1_v2 ] ) , @xmath112 generally take different values ( but they are expected to be the same order ) .
the transverse spin correlation also contributes to @xmath0 , but it must be a thermal activation type @xmath113 , in which @xmath114 is proportional to the binding energy of @xmath12 magnons , that is , the single - magnon gap .
therefore , it can be negligible when temperature is sufficiently smaller than @xmath115 .
the absence of the contribution from the transverse spin correlation is a characteristic feature of the multipolar tl liquids . the temperature dependence of @xmath0 given in eq .
( [ eq:1/t1_multipolar ] ) is consistent with the previous prediction in ref . .
the nmr relaxation rate @xmath0 diverges with lowering temperature in the sdw regime ( @xmath98 ) , whereas it decays algebraically with lowering @xmath22 in the high - field multipolar regime ( @xmath97 ) .
thus the temperature dependence changes significantly at the crossover point @xmath99 .
this feature contrasts sharply with the result of usual tl liquids ( see the next subsection ) .
more detailed properties of @xmath0 will be discussed in sec .
[ sec : h - t - dep ] . here
, we quickly summarize the well - known effective theory for the spin-@xmath1 af heisenberg chain in magnetic field to compare with the previously mentioned theory for multipolar liquids .
the hamiltonian is given by @xmath117 where @xmath118 is af exchange coupling .
as is well known , the low - energy physics of this model belongs to the tl - liquid universality class from zero field to the saturation field @xmath119 .
the low - energy hamiltonian has the same free - boson form as eq .
( [ eq : eff_multipolar ] ) : @xmath120,\end{aligned}\ ] ] where @xmath121 and @xmath122 are the pair of scalar fields . the tl - liquid parameter @xmath123 and the velocity @xmath124
are exactly determined from the bethe ansatz.@xcite in this model , @xmath123 changes monotonically from 1/2 to 1 when @xmath10 increases from zero to @xmath56 .
similarly to the multipolar tl liquids , the value of @xmath124 is of order of @xmath125 except for the saturation limit @xmath74 where @xmath126 .
the tl - liquid theory is valid only for @xmath127 .
spin operators are bosonized as @xmath128.\end{aligned}\ ] ] here @xmath129 , @xmath130 , and @xmath131 are non - universal coefficients , whose values are numerically determined as a function of @xmath10 ( or @xmath132 ) in ref . .
equations ( [ eq : eff_tl ] ) and ( [ eq : spin_boson ] ) enable us to estimate @xmath0 .
the result is @xmath133 where @xmath109 and @xmath134 .
the first and second terms on the right - hand side originate from the longitudinal spin correlation , while the third term originates from the transverse spin correlation .
in contrast to the multipolar liquids , both longitudinal and transverse spin correlations yield power - law - type functions of temperature .
equation ( [ eq:1/t_1_tl ] ) shows that either @xmath135 or @xmath136 term diverges at low temperatures regardless of the value of @xmath123 , except for @xmath137.@xcite this diverging behavior of @xmath0 with decreasing temperature is a common property of usual tl liquids in 1d magnets , and is quite different from the decaying behavior of @xmath0 in the multipolar liquid phases .
utilizing the formulas ( [ eq:1/t1_multipolar ] ) and ( [ eq:1/t_1_tl ] ) , we examine the magnetic - field and temperature dependence of @xmath0 in both multipolar and usual tl liquid phases . among the three multipolar phases , we concentrate on the quadrupolar tl liquid , which is most widely expanded in the @xmath139-@xmath10 space and is believed to be realized in several real magnets .
we show that @xmath0 in the quadrupolar tl liquid exhibits characteristic features distinct from that in the usual tl liquid . in eqs .
( [ eq:1/t1_multipolar ] ) and ( [ eq:1/t_1_tl ] ) , the values of hyperfine constants @xmath140 depend on crystal structure , the spatial relationship between electron and nuclear spins , and the direction of applied field @xmath10 . for simplicity
, we assume that all the constants @xmath140 are equal throughout this section .
let us first consider @xmath0 of the quadrupolar liquid phase .
we can evaluate the expression of @xmath0 given in eq .
( [ eq:1/t1_multipolar ] ) , once the parameters @xmath71 , @xmath72 , and @xmath141 are numerically estimated . to this end
, we perform numerical calculations of spin-@xmath28 @xmath2-@xmath3 chains using the dmrg method . for @xmath71 and @xmath141 , we compute the equal - time longitudinal spin correlation function @xmath142 and the spin polarization @xmath143 in the ground state .
we then fit the data to an analytic form obtained by the bosonization method to estimate @xmath71 and @xmath141 .
( the details of the method have been presented in refs . , , and . ) for the velocity @xmath72 , we use uniform magnetic susceptibility . from the bosonization formula ( [ eq : boson_phase ] ) , the susceptibility is given by @xmath144 in the @xmath12-th order multipolar phase .
we can thus obtain the value of @xmath72 from the magnetization curve ( i.e. , susceptibility ) determined by the dmrg calculation .
we determine the values of @xmath71 , @xmath72 , and @xmath141 at @xmath145 , 0.1 , 0.15 , @xmath146 , 0.45 for several values of @xmath139 .
we have found that the tl - liquid parameter @xmath71 is basically an increasing function of the field @xmath10 ( see figs .
14 and 18 of ref . ) , while @xmath141 has a fairly small field dependence .
as expected , @xmath147 is shown to be of order of the 1d couplings @xmath2 and @xmath3 for intermediate magnetization and approaches zero near the saturation .
for example , in the case of @xmath148 ( @xmath149 ) , @xmath150 ( @xmath151 ) at @xmath152 , while @xmath153 ( @xmath154 ) at @xmath155 . in the magnetic quadrupolar tl - liquid phase in the @xmath2-@xmath3 spin chain ( [ eq : zigzag ] ) for ( a ) @xmath156 and ( b ) @xmath157 .
left panels show the results for the whole range of field @xmath158 while right panels show the same data in the vicinity of the saturation field in an enlarged scale .
solid symbols represent the data obtained from eq .
( [ eq:1/t1_multipolar ] ) with numerical values of @xmath71 , @xmath72 , and @xmath141 .
lines connecting them are guide for the eyes .
the vertical dashed lines represent the saturation field and the crossover line between the sdw ( @xmath159 ) and nematic ( @xmath160 ) regimes . in ( b ) , the boundary to the low - field vector - chiral phase is shown by vertical solid line .
, width=325 ] figure [ fig : nmr_nematic ] shows the magnetic - field dependence of @xmath0 in the @xmath2-@xmath3 spin chain for @xmath161 and @xmath149 at several values of temperature . except for the vicinity of the saturation , @xmath0 decreases with increasing @xmath10 ( or @xmath132 ) for sufficiently low temperatures @xmath162 .
this is mainly due to the monotonic field dependence of the tl - liquid parameter @xmath71 , and is a distinct feature of the multipolar tl liquids , in contrast with that of the spin-@xmath1 af chain ( see the next paragraph ) .
this feature becomes less significant as temperature increases , and it becomes almost independent of the field at relatively high temperatures @xmath163 . on the other hand , in a narrow region near the saturation field , @xmath0 increases rapidly with @xmath10 since the velocity @xmath72 approaches zero .
we also confirm the previous prediction@xcite on the temperature dependence of @xmath0 : @xmath0 increases divergently with lowering @xmath22 in the low - field sdw regime ( @xmath159 ) , whereas it decreases algebraically in the high - field quadrupolar regime ( @xmath160 ) .
explicit temperature dependence is presented in fig .
[ fig : nmr_nematic_2 ] . in the magnetic quadrupolar tl - liquid phase in the @xmath2-@xmath3 spin chain ( [ eq : zigzag ] ) for ( a ) @xmath156 and ( b ) @xmath157 and several fixed values of magnetization @xmath132 .
, title="fig:",width=245 ] in the magnetic quadrupolar tl - liquid phase in the @xmath2-@xmath3 spin chain ( [ eq : zigzag ] ) for ( a ) @xmath156 and ( b ) @xmath157 and several fixed values of magnetization @xmath132 . ,
title="fig:",width=245 ] the characteristic behavior in the low - temperature regime @xmath164 is consistent with the previous statement .
next , we compare these features of the relaxation rate in the nematic liquid with those in a usual tl liquid of the spin-@xmath1 af heisenberg chain ( [ eq : afchain ] ) . similarly to the nematic case , we can evaluate @xmath0 from the formula ( [ eq:1/t_1_tl ] ) if values for the parameters @xmath123 , @xmath124 , @xmath129 , and @xmath130 are prepared . as mentioned in sec .
[ sec : multipolar_2_tl ] , @xmath123 and @xmath124 are exactly determined by the bethe ansatz@xcite and the coefficients @xmath129 and @xmath130 have been evaluated numerically.@xcite using these accurate values , we numerically determine @xmath0 of the model ( [ eq : afchain ] ) .
the result is shown in fig .
[ fig : nmr_usualtl ] .
it indicates that @xmath0 is a monotonically increasing ( decreasing ) function of @xmath10 ( @xmath22 ) in the tl - liquid phase of the model ( [ eq : afchain ] ) for sufficiently low temperatures .
these two properties are completely different from those of the nematic tl - liquid phase shown in figs .
[ fig : nmr_nematic ] and [ fig : nmr_nematic_2 ] .
the relaxation rate of real quasi-1d spin-@xmath1 af magnets has been observed in somewhat restricted regime of @xmath22 and @xmath10 , for example , in refs . .
their results seem to agree nicely with our results in fig .
[ fig : nmr_usualtl ] . for the standard tl - liquid phase in the spin-@xmath1 heisenberg chain ( [ eq : afchain ] ) ; ( a ) field dependences for fixed temperatures and ( b ) temperature dependences for fixed magnetization . , title="fig:",width=245 ] for the standard tl - liquid phase in the spin-@xmath1 heisenberg chain ( [ eq : afchain ] )
; ( a ) field dependences for fixed temperatures and ( b ) temperature dependences for fixed magnetization . ,
title="fig:",width=245 ] finally , let us separately estimate each term in the expressions of @xmath0 , eqs .
( [ eq:1/t1_multipolar ] ) and ( [ eq:1/t_1_tl ] ) .
they are depicted in fig .
[ fig : nmr_each ] . for ( a ) the magnetic quadrupolar tl - liquid phase in the @xmath2-@xmath3 spin chain ( [ eq
: zigzag ] ) with @xmath156 and ( b ) the standard tl - liquid phase in the spin-@xmath1 heisenberg chain ( [ eq : afchain ] ) .
left ( right ) panels show the data for @xmath165 and @xmath166 ( @xmath167 and @xmath168 ) .
in addition to the total value of @xmath0 , each term in eqs .
( [ eq:1/t1_multipolar ] ) and ( [ eq:1/t_1_tl ] ) is shown separately .
, width=325 ] we find that in the heisenberg chain ( [ eq : afchain ] ) the relaxation rate @xmath0 is clearly governed by the contribution from the transverse spin correlation , that is , the @xmath136 term . in the nematic case ,
in which the transverse spin fluctuation is gapped , the @xmath135 term becomes dominant in a wide magnetic - field range and is responsible for the behavior of @xmath0 decreasing with field @xmath10 .
this figure definitely demonstrates that characteristic features of @xmath0 in the @xmath2-@xmath3 spin chain originates from the absence of gapless modes in the transverse spin excitations . from these results
, we conclude that the nmr relaxation rates @xmath0 in the nematic and standard tl - liquid phases exhibit quite different behavior in @xmath22-@xmath10 space , and hence we propose that its @xmath22 and @xmath10 dependence can be used to distinguish the nematic liquid phase from the standard tl liquid .
similar characteristic behavior of @xmath0 is also expected in the higher - order multipolar liquid phases in the @xmath2-@xmath3 spin chain , since these phases share essentially the same properties of spin correlations and the field @xmath10 dependence of the tl - liquid parameter .
furthermore , a nematic tl - liquid phase in the sdw regime ( @xmath97 ) is also shown to appear in a wide parameter region of the @xmath2-@xmath3 chain with af @xmath2 .
@xcite this af-@xmath2 sdw@xmath20 phase , where the transverse spin correlation is short - ranged , is also expected to show a peculiar @xmath10 dependence of @xmath0 , which is distinct from that in the usual tl liquid .
so far we have discussed low - temperature behavior in spin isotropic pure 1d systems . in this section
, we consider the temperature dependence of @xmath0 in wide temperature range , taking interchain couplings into account .
we also discuss the relation between the form of hyperfine coupling tensor @xmath169 and the direction of field @xmath10 , as well as effects of dzyaloshinsky - moriya ( dm ) interactions .
first we consider a wide temperature window including higher and lower temperatures in which simple effective hamiltonians ( [ eq : eff_multipolar ] ) and ( [ eq : eff_tl ] ) are no longer valid for describing the physics of real quasi-1d magnets . at sufficiently low temperature , which is lower than the energy scale of weak 3d interchain couplings @xmath170
, long - range - ordered phases usually emerge due to 3d couplings and hence 1d effective theories can not be applicable . in the case of the sdw regime , @xmath0 exhibits a divergence with a critical exponent near the critical temperature @xmath171 , if 3d ordering occurs through a continuous ( second - order ) transition . in case of nematic regime ,
the nematic ordering at finite temperature may induce singularity in @xmath0 . on the other hand , at high temperature ,
the effective tl - liquid theory becomes unreliable when the energy scale @xmath172 is increased up to the order of @xmath147 , which is of the order of the 1d couplings @xmath2 and @xmath3 , except for the saturation limit @xmath74 where @xmath76 .
deviation from the tl - liquid theory also comes from breaking of magnon bound states at high temperatures .
the binding energy @xcite is numerically estimated , at most , as @xmath173 and @xmath174 for @xmath175 , and @xmath176 , respectively , which gradually decreases with lowering magnetization and vanishes at the boarder to the lower - field vector chiral phase .
thus , our prediction is valid in the temperature range @xmath177 $ ] . in the higher - temperature regime @xmath178
, the relaxation rate converges to a constant value.@xcite we draw schematic patterns of the temperature dependence of @xmath0 in the nematic tl liquid in figs .
[ fig : several_1/t_1](a ) and ( b ) . of the nematic tl liquid in wide temperature region for several parameter settings . in the regime @xmath179 , our prediction based on low - energy effective theories
can be applicable.,width=264 ] next we discuss the anisotropic case @xmath180 ( we have assumed @xmath181 so far for simplicity ) . in the extreme case @xmath182 , we can no longer neglect the contribution from the exponentially decaying transverse spin correlation in @xmath0 . in this extreme case ,
power - law behavior in @xmath0 of multipolar tl liquids is negligible down to extremely low temperature , hence we observe a thermal activation form @xmath113 in @xmath0 in a low - temperature regime @xmath183 .
thus it is easy to detect a characteristic feature of multipolar tl liquids .
if target magnets have sufficiently high crystallographic symmetry , the principal axes for hyperfine coupling tensor @xmath46 can be defined and the tensor is diagonalized . in this case , tuning the field direction parallel to a principal axis , we can eliminate off - diagonal elements @xmath184 and @xmath185 , and equivalently set @xmath186 .
therefore , the setup of @xmath10 parallel to the axis offers an easy way of measuring the transverse spin gap and distinguishing multipolar liquids from ordinary tl liquid .
figures [ fig : several_1/t_1](c ) and [ fig : several_1/t_1](d ) present the schematic temperature dependence of @xmath0 in the nematic and sdw@xmath20 tl liquids for the anisotropic case @xmath182 . in the opposite limit @xmath187 , our predictions in figs .
[ fig : nmr_nematic ] and [ fig : nmr_nematic_2 ] are highly reliable in @xmath188 .
however , we should note that in this limit @xmath187 , @xmath0 becomes insensitive to the disappearance of the algebraically - decaying transverse spin correlation , the hallmark of the nematic liquid , and is not efficient in distinguishing the usual and nematic tl liquids . in the rest of this section ,
we consider the effects of magnetic anisotropies which generally exist in real magnets . in spin-@xmath1 systems ,
one of the most realistic anisotropies is the dm interaction defined by @xmath189 for @xmath190 .
possible values of @xmath191 , @xmath192 , and the dm vector @xmath193 depend strongly on the crystal structure of each compound . since the dm interaction can sometimes induce an excitation gap accompanied with local spin polarization , it may violate the multipolar tl liquids .
we discuss the effects of the dm interaction on the nematic tl liquid in the low - energy theory . in the weak - coupling theory@xcite for the spin-@xmath28 @xmath2-@xmath3 chain with @xmath194 ,
the nematic phase is expressed by the effective hamiltonian @xmath195\nonumber\\ & & + c_1 ( 2a)^{-1}\sin(\sqrt{8\pi}\phi_-+\pi m ) \nonumber\\ & & + c_2 ( \partial_y\theta_+)\sin(\sqrt{2\pi}\theta_- ) + \cdots\end{aligned}\ ] ] under the condition that only the @xmath196 term is relevant to the gaussian model , where @xmath197 , @xmath198 , and @xmath199 is a boson - field pair defined in each of the decoupled af-@xmath3 chains ( see sec .
[ sec : multipolar_2_tl ] ) .
the coupling constants @xmath200 are proportional to @xmath2 .
the @xmath196 term makes @xmath201 pinned . as a result
, the transverse spin correlation decays in an exponential fashion .
the remaining @xmath202 sector induces the gapless behavior of longitudinal - spin and nematic correlation functions . on the other hand ,
the @xmath203 term is known to induce a vector chiral phase @xcite with @xmath204 in a lower - field regime .
@xcite let us concentrate on the former nematic tl liquid in the following .
using the effective theory ( [ eq : eff_zigzag_2 ] ) , we investigate five typical situations : @xmath205 , @xmath206 , @xmath207 , @xmath208 , and @xmath209 .
let us first consider the case of @xmath210 , that is , the dm vector is parallel to the applied field @xmath10 . in this case , we find that the bosonized dm couplings @xmath211 contain slowly - moving bosonic terms without oscillating factors @xmath66 or @xmath212 ( @xmath213 is an integer ) . they are represented as [ eq : dm_boson_dz ] @xmath214\nonumber\\ & & -\sin(\sqrt{2\pi}\theta_-)\big[(2a)^2\frac{\pi}{2 } \big((\partial_y\theta_+)^2\nonumber\\ & & + ( \partial_y\theta_-)^2\big)+\cdots\big]\big\},\\ { \cal h}_{dm}^{\rm{\rnum 3}}&\sim & d_z\sqrt{\pi}(2a)\partial_y\theta_++\cdots , \\ { \cal h}_{dm}^{\rm{\rnum 4}}&\sim & d_z\sqrt{\pi}(2a)\partial_y\theta_-+\cdots.\end{aligned}\ ] ] in general , dm terms with @xmath210 are invariant under a global u(1 ) spin rotation @xmath215 , which corresponds to a shift of the phase field @xmath216 .
therefore , the bosonized dm terms with @xmath217 do not involve any vertex operator with @xmath218 .
in addition , all the translationally - symmetric dm terms do not contain any vertex term with @xmath219 at least for the case of incommensurate values of @xmath132 , since the @xmath220-site translation induces @xmath221 and equivalently @xmath222 .
equation ( [ eq : dm_boson_dz ] ) is consistent with these symmetry arguments .
two dm interactions @xmath223 tend to lock @xmath224 together with the @xmath203 term in eq .
( [ eq : eff_zigzag_2 ] ) , both of which compete with the @xmath196 term . in the nematic phase , we know that the @xmath196 term is most relevant and hence the dm interactions @xmath223 with sufficiently small @xmath225 are negligible . in the strong dm coupling regime , these dm terms compete with the @xmath196 term and could change the nematic liquid into a chiral phase with @xmath226 . on the other hand ,
the leading terms of @xmath227 can be absorbed into the free - boson part by shifting the field @xmath228 .
the shift brings about a finite expectation value @xmath229 , but we note that these chiralities do not accompany any spontaneous symmetry breaking .
the shift also affects the form of spin dynamical structure factors , that is , the gapless points of @xmath105 are slightly changed and a small asymmetry of @xmath42 dependence emerges .
in contrast to @xmath223 , @xmath227 do not compete with the @xmath196 term .
therefore , chirality and nematic quasi long - range order can coexist in the latter case of @xmath227 . this symmetry argument indicates that the dm terms with @xmath217 do not introduce any vertex operator in the @xmath202 sector , and this statement would be true in the wide nematic - liquid region regardless of the value of @xmath2 .
the gapless nature of the @xmath202 sector hence survives even after introducing these dm terms . from these discussions , we conclude that the nematic tl liquid survives even in the presence of dm terms with @xmath210 at least when @xmath230 .
thus , the nematic liquid is relatively stable against dm interactions with @xmath210 .
for the case of @xmath231 , the u(1 ) rotational symmetry is broken and the emergence of vertex operators @xmath232 is generally allowed .
in fact , @xmath233 is bosonized as @xmath234 this leading term has scaling dimension @xmath235 and can generate a staggered magnetization along the @xmath236 axis in each af-@xmath3 chain . in the weak @xmath2 limit , @xmath233 is more relevant than the @xmath196 term with scaling dimension @xmath237 and it can violate the nematic liquid phase . [ in this limit , @xmath238 approaches the value of original tl - liquid parameter @xmath123 ( @xmath239 ) of the af-@xmath3 chain in magnetic field .
] when @xmath2 is sufficiently strong , the @xmath196 terms defeats the perturbation @xmath233 and protects the nematic liquid phase , pinning @xmath201 . in this case
, @xmath240 generates new operators @xmath241 and @xmath242 via the renormalization - group process . the first term with scaling dimension @xmath243 can open a gap in the @xmath202 sector and induce a transverse staggered magnetization , since the u(1 ) spin symmetry is broken by the dm term . in the weak dm coupling regime , thus , this dm term perturbatively deforms nematic spin liquid , opening a small gap in longitudinal spin modes and inducing a small expectation value of transverse spins .
the other dm interactions @xmath244 with @xmath193 parallel to the @xmath245 axis do not possess any slowly - moving operator within a naive calculation based on the bosonization , but they generally have the ability to generate @xmath232 , which is allowed from the symmetry argument .
thus , the gapless nature of the @xmath202 sector is expected to be fragile and unstable against dm terms with @xmath246 .
these discussions on dm terms indicate that if we apply magnetic field parallel to the dm vector @xmath193 , the nematic tl liquid is stably realized and our prediction of @xmath0 is reliable in wider @xmath10-@xmath22 space .
even when @xmath193 is not parallel to @xmath10 , the nematic phase will also survive if the dm coupling is small , @xmath247 , and if the system is in the temperature regime @xmath248 .
we have evaluated accurately the field and temperature dependence of the nmr relaxation rate @xmath0 in magnetic quadrupolar ( spin nematic ) tl liquid of the spin-@xmath1 @xmath2-@xmath3 chain , combining field - theoretical techniques with dmrg results ( see figs . [
fig : nmr_nematic ] and [ fig : nmr_nematic_2 ] ) . as a comparison ,
we have also calculated @xmath0 of the spin-@xmath1 af heisenberg chain , using field theories , the bethe ansatz , and dmrg method ( see fig .
[ fig : nmr_usualtl ] ) . in the nematic and sdw@xmath20 tl - liquid phase at a low temperature @xmath249 , the relaxation rate @xmath0 first decreases with increasing magnetic field and then rapidly increases near saturation . in the higher - temperature regime @xmath163 ,
the field dependence of @xmath0 becomes quite small except in the vicinity of saturation .
the decreasing behavior of @xmath0 with increasing @xmath10 comes from the monotonic increase in the tl - liquid parameter @xmath71 , while the rapid increase in @xmath0 near saturation is attributed to the decrease of velocity @xmath250 .
the monotonic property of @xmath71 is also essential for the characteristic temperature dependence@xmath251 of @xmath0 : with lowering temperature , @xmath0 increases in an algebraic form in the lower - field sdw regime ( @xmath98 ) , whereas it decreases in the higher - field nematic regime ( @xmath97 ) .
these characteristic @xmath10 and @xmath22 dependences could be a signature of the nematic and sdw@xmath20 tl - liquid phase .
similar features are also expected to appear in higher - order multipolar tl liquids , for example , octupolar and hexadecapolar tl liquids .
probing the @xmath10 dependence would be easier than doing the @xmath22 dependence since the former does not necessarily require the accession to high - field regime , where nmr measurements are difficult .
a combination of the @xmath10 and @xmath22 dependences of @xmath0 and the gapless behavior observed from bulk quantities ( such as specific heat and magnetic susceptibilities ) would present indirect but strong evidence for multipolar tl - liquid phases . in sec .
[ sec : realmagnets ] , we also considered the effects of spin anisotropies and interchain couplings , which are neglected in the ideal @xmath2-@xmath3 spin chain model .
in particular , we point out that if the direction of the external field @xmath10 can be parallel to the principal axis of the hyperfine - coupling tensor , all the algebraic contributions in the temperature dependence of @xmath0 , that is , @xmath252 terms , disappear . as a result ,
@xmath0 of multipolar phases in the @xmath2-@xmath3 chain becomes a thermal activation form @xmath113 .
in addition , we predict that the nematic tl liquid is stable for small dm terms with the dm vector parallel to the applied field , while it can be easily deformed by the dm terms , thereby accompanied with small transverse staggered magnetization , when the dm vector is perpendicular to the field .
this indicates that we should apply magnetic field parallel to the dm vector to obtain a stable nematic liquid phase in real quasi-1d @xmath2-@xmath3 magnets .
finally we comment on some real compounds . recently ,
quasi-1d edge - sharing cuprate magnets , for example , @xmath26 , @xmath4 , @xmath24 and @xmath25 , have been studied extensively as low - dimensional frustrated or multiferroic systems .
their magnetic properties are believed to be described by spin-@xmath1 @xmath2-@xmath3 chains in a certain temperature regime . except for @xmath24 , a 3d ordering has been observed below a very low critical temperature @xmath253 at least in zero magnetic field .
furthermore , the values of @xmath2 and @xmath3 have been semi - quantitatively estimated in several ways .
the estimated coupling constants and the critical temperatures in @xmath26 ( refs . ) , @xmath25 ( refs . ) , and @xmath24 ( ref . ) are , respectively , @xmath254 k , 86 k , 24 k ) , @xmath255 k , 21 k , 2.8 k ) , and @xmath256 k , 51 k ) , while two different results have been reported for @xmath4 ( refs . ) ; @xmath257 k , 45 k , 2 k ) and @xmath258 k , 91 k , 2 k ) .
the critical temperatures @xmath253 are thus small compared to the magnitude of @xmath259 except for @xmath26 .
it means that three dimensionality is small in these compounds .
we note that our prediction for @xmath0 could be applied to the temperature condition @xmath260 . among these compounds ,
the magnetization process of @xmath4 has been intensively studied in some experimental groups . in the low - field regime including zero field , spiral phases exist at low temperatures . above @xmath261
t , the spiral phase turns into another phase which was concluded by nmr measurements@xcite to be a modulated collinear phase .
the appearance of both the spiral and modulated collinear phases was well understood@xcite as a consequence of the vector chiral phase and the incommensurate sdw@xmath20 phase in the 1d @xmath2-@xmath3 spin chain .
in addition , quite recently , a new phase transition has been observed @xcite at @xmath262 t , where the saturation field is @xmath263 t. comparing the result with the phase diagram in the 1d @xmath2-@xmath3 spin chain,@xcite we expect the new high - field phase for @xmath264 t to be a nematic long - range ordered phase .
so far , the magnetic structure of this new phase has not been experimentally identified at all .
we expect that nmr measurements above the critical temperatures of this new high - field phase and of the intermediate - field modulated collinear phase would be useful to verify whether the new phase is a nematic phase .
we also note that @xmath25 has a rather small saturation field @xmath265 t. @xcite this might be an ideal material for measuring the @xmath10 dependence of @xmath0 up to the saturation field .
we thank akira furusaki , masayuki hagiwara , tetsuaki itou , anja wolter , and yukio yasui for fruitful discussions .
this work was supported by grants - in - aid for scientific research from mext , japan ( grants no . 21740277 , no . 21740295 , no .
22014016 ) .
m. enderle , c. mukherjee , b. fk , r. k. kremer , j .-
broto , h. rosner , s .- l .
drechsler , j. richter , j. malek , a. prokofiev , w. assmus , s. pujol , j .- l .
raggazzoni , h. rakoto , m. rheinstdter , and h. m. rnnow , europhys . lett . * 70 * , 237 ( 2005 ) . in the narrow range
@xmath266 , the soft two - magnon bound state has an incommensurate momentum @xmath267 to realize a two - component tl liquid . in this paper
, we do not consider this case and focus on the one - component mulitpolar tl liquids of the multimagnon bound states with momentum @xmath58 .
however , it is naively expected that the incommensurate multipolar liquid qualitatively shares the features of @xmath0 in the commensurate multipolar liquids ( see sections iii and iv ) which mainly result from the formation of multimagnon bound states
. there are logarithmic corrections to eq .
( [ eq:1/t_1_tl ] ) near the zero - field su(2)-symmetric point with @xmath137 , which we omitted .
see s. sachdev , phys .
b * 50 * , 13006 ( 1994 ) ; m. takigawa , n. motoyama , h. eisaki , and s. uchida , phys .
lett . * 76 * , 4612 ( 1996 ) ; m. takigawa , o.a .
starykh , a.w .
sandvik , and r.r.p .
singh , phys .
b * 56 * , 13681 ( 1997 ) .
the magnon binding energy is defined as follows .
first we prepare @xmath268 , where @xmath269 is the lowest energy eigenvalue in the subspace with fixed @xmath270 ( @xmath271 is the total number of sites ) . calculating the binding energies of the two definitions @xmath272 and @xmath273
, we obtain the true binding energy from the minimum value of them , @xmath274 . | it is generally difficult to experimentally distinguish magnetic multipolar orders in spin systems .
recently , it was proposed that the temperature dependence of the nuclear magnetic resonance relaxation rate @xmath0 can involve an indirect , but clear signature of the field - induced spin nematic or multipolar tomonaga - luttinger ( tl ) liquid phase [ phys .
rev .
b * 79 * , 060406(r ) ( 2009 ) ] . in this paper , we evaluate accurately the field and temperature dependence of @xmath0 in spin-@xmath1 frustrated @xmath2-@xmath3 chains combining field - theoretical techniques with numerical data .
our results demonstrate that isotherms of @xmath0 as a function of magnetic field also exhibit distinctive non - monotonic behavior in spin nematic tl liquid , in contrast with the standard tl liquid in the spin-@xmath1 heisenberg chain .
the relevance of our results to quasi one - dimensional edge - sharing cuprate magnets , such as @xmath4 , is discussed . |
the littleness of the planck length is indescribable .
the immense size the observable universe is paralyzing .
however , the geometric mean of the two characteristic values results in a tangible quantity .
explicitly , @xmath7 , where the planck length @xmath8 m and the gigaparsec @xmath9 m. as we will show in the example discussed in the essay , better understanding of both the super - short and super - large scales can be gained thanks to this property .
so , how does it work ?
there are two main possibilities .
the first one is a result of accumulation of the tiny planck scale deviations acting on particles propagating across cosmological distances .
the second possibility exploits the evolution of the universe .
the universe is as big as it is because it is expanding . as a result ,
the planck scale in the early universe , multiplied by the growth factor becomes macroscopic in the mature universe . in both cases
, the multiplication law brings the planck scale closer to the scales of our perception .
we do not know yet what the planck scale physics is .
however , by taking the use of the compensation of scales discussed above , one may already try to falsify the theoretical models of quantum gravitational phenomena , which we have .
the theory of planck scale physics which we are going to examine here is causal dynamical triangulations ( cdt ) @xcite .
the theory is a non - perturbative approach based on the path integral formulation of quantum mechanics .
the crucial ingredient of the theory is a causality condition imposed at the level of wick - rotated time coordinate . in similarity to statistical physics , the different space - time configurations are evaluated using monte carlo computer simulations .
various promising results arose from numerical studies of the quantum gravitational system with the positive cosmological constant .
one of the most profound is the emergence of four dimensional de sitter space - time @xcite in the so - called phase c. however , as reflected by features of the process of diffusion taking place at the simplicial manifold of cdt , the reconstructed de sitter space - time is not fully classical .
in particular , it has been shown that the spectral dimension characterizing the diffusion process can be parametrized in the following way @xcite : @xmath10 where @xmath11 is the diffusion time . at the representative point in phase c ( @xmath12 , @xmath13 ) the values @xmath14 , @xmath15 and @xmath16 @xcite and @xmath17 , @xmath18 and @xmath19 @xcite
have been obtained from numerical simulations . in both cases , four dimensional space - time ( @xmath20 )
is correctly recovered in the ir limit ( @xmath21 ) .
however , in the uv limit ( @xmath22 ) dimensional reduction to @xmath23 is predicted .
this effect is not reserved to cdt only and has been noticed in other approaches to quantum gravity as well @xcite . the expression ( [ spectralparametrization ] ) is a starting point for further considerations .
after fixing the ir value @xmath24 to be precisely equal to the topological space - time dimension @xmath25 and parametrizing the departure from @xmath26 in uv by @xmath27 we obtain @xmath28 such that @xmath29 .
the parameter @xmath30 can be related to the energy scale of the dimensional reduction : @xmath31 .
the spectral dimension discussed in the previous section is formally defined as follows @xmath32 where @xmath33 is the average return probability of the random walk ( diffusion ) process . here
, @xmath34 is the invariant measure in the momentum space , which we assume to be the classical one @xmath35 .
@xmath36 is the laplace operator in momentum space . in the case of the euclidean 4-dimensional space @xmath37
this expression might be viewed as the wick rotated @xmath38 version of the analogous formula in the minkowski space @xmath39 , which defines dispersion relation for the free fields living on the space - time ( @xmath40 ) . in this case , the equations ( [ spectraldim ] ) and ( [ returnprobability ] ) predict @xmath41 , where @xmath25 is the topological dimension .
cdt predicts that the spectral dimension varies as a function of the diffusion time @xmath11 .
the corresponding moment laplace operator has to be , therefore , deformed with respect to the classical case . here , we parametrize this departure in the following manner @xmath42 where @xmath43 is some unknown function of @xmath44
. the task would be now to recover the form of the function @xmath43 for the cdt spectral dimension ( [ cdtsd2 ] ) . in this case , the expression for the return probability is @xmath45 where integrations over @xmath46 and the angular part of @xmath47 have been performed . as it has been shown in ref .
@xcite , the relation ( [ rpreduced ] ) can be converted into the form of the inverse laplace transform @xmath48 the @xmath49 has to satisfy the condition @xmath50 , where @xmath51 denotes singularities of @xmath52 .
the @xmath53 function is recovered by integrating definition ( [ spectraldim ] ) with use of parametrization ( [ cdtsd2 ] ) .
we find that @xmath54 which , when applied to ( [ inverselaplace ] ) , leads to @xmath55 where @xmath56 is the confluent hypergeometric function .
equation ( [ defdisprel ] ) , is an entangled form of the deformed dispersion relation @xmath57 .
the function @xmath58 for @xmath59 is shown in fig .
[ disp ] . in the ir limit
@xmath60 the dispersion relation is approximated by @xmath61 so the classical limit is correctly recovered .
however , when passing to the high energy range ( the uv limit ) the dispersion relation undergoes deflection to the form @xmath62 both of the obtained limiting behaviors of the deformed dispersion relation will now be used to construct the relation of cdt to astrophysical and cosmological observations .
the dispersion relation derived in the previous section can now be used to find the group velocity of photons . for the particles having energy @xmath63 the ir limit approximation ( [ irlim ] ) may be applied , leading to the quadratic deviation @xmath64 this expression predicts that the group velocity is increasing with the energy ( superluminal behavior ) .
therefore , when a bunch of photons of different energies are emitted simultaneously the ones with the highest energy will arrive first to the observer .
this effect can be quantified by the arrival time difference between @xmath46 and @xmath65 energy photons . with use of formula ( [ groupvelocity ] ) , the corresponding expression is @xmath66 , where @xmath67 is the distance to a source .
although a huge value of the energy scale @xmath68 is expectedgev . ] , the multiplication by a sufficiently large distance @xmath67 may bring the value of @xmath69 close to the observational window .
the effect can be constrained with use of the signals from the high - energy astrophysical sources such as gamma ray bursts ( grb ) @xcite .
in particular , the grb 090510 remote by @xmath70 mpc may be used . using the constraint from the _ fermi_-large area telescope @xcite one can find that the energy scale of the dimensional reduction is @xmath71 gev at ( 95 @xmath1 cl ) .
this constraint is definitely much stronger than any obtained from the accelerator physics experiments ( which are reaching energies of the order of @xmath72 gev ) .
however , because of the quadratic ( in energy ) from of the effect , the constraint is distant from the planck scale .
because of the quadratic nature of the ir variation of the group velocity , perhaps the more promising is an application of dimensional reduction to cosmology . here , instead of the ir , the uv part of the dispersion relation will play a crucial role .
the potential relevance of the uv scaling of the modified dispersion relation in cosmology has been suggested in ref .
it has been shown that dimensional reduction to @xmath73 leads to the scale invariant vacuum fluctuations of the primordial perturbations in the uv domain .
what has not been shown is how the shape of the spectrum changes when we try to relate it with the one that is observationally relevant . here
, we show that a deviation from @xmath73 leads to a tilt in the power spectrum .
this prediction is confronted with the observations of the cosmic microwave background radiation ( cmb ) .
the resulting cdt dispersion relation @xmath57 can be incorporated into the equation of modes for the cosmological perturbations in the following way @xmath74 here , @xmath75
( @xmath76 is the scale factor ) and the classical limit @xmath77 is recovered correctly for @xmath78 .
the value of @xmath79 depends on which kind of perturbations is considered . for the gravitational waves @xmath80 and for the scalar perturbations @xmath81 . in the sub - hubble limit @xmath82
, the vacuum normalization of the @xmath83 mode functions leads to @xmath84 applying this to the definition of the scalar power spectrum we find that for @xmath85 @xmath86 where @xmath87 .
furthermore , we assume that the universe is filled with barotropic matter : @xmath88 , with @xmath89 . indeed , for @xmath73 the power spectrum ( [ scalarps ] ) becomes scale invariant @xmath90 . while it is tempting to relate the obtained spectrum with that inferred from the cmb nearly scale - invariant spectrum of the primordial perturbations , it is not allowed to do so directly .
the spectrum ( [ scalarps ] ) corresponds to the uv regime ( @xmath85 ) which has to be suitably converted into the super - hubble spectrum probed observationally . in general
, this would require evolving the modes starting from the uv domain until crossing the hubble horizon .
this will unavoidably affect the shape of the power spectrum .
a scenario in which knowledge of the intermediate evolution is not required corresponds to the case when the hubble horizon is located in the uv domain . in that case _ freezing _ of the modes occurs already at the level of the uv spectrum ( [ scalarps ] ) . in this particular situation
the resulting tensor - to - scalar ratio @xmath91 can be directly related to the cmb data ) . ] .
the best current observational bound at the pivot scale @xmath92 is @xmath93 ( bicep2/_keck array _ and _ planck _ ) @xcite , which implicates that @xmath94 when the perturbations were formed .
the universe had to , therefore , be very close to the de sitter phase when the perturbations were formed .
cosmic inflation is still needed !
the value of the power spectrum accessible observationally is at the horizon - crossing . in our case
, the horizon - crossing condition @xmath95 translates into @xmath96 @xmath97 . ] for @xmath98 .
using this , the spectral index of scalar perturbations at the horizon - crossing is @xmath99 the scale invariance is still recovered for @xmath100 .
however , by comparing the prediction with the _ planck _ satellite measurement of the spectral index ( @xmath101 ) @xcite and the constraint on the tensor - to - scalar ratio @xcite we find that @xmath102 the value is in overt contradiction with the cdt - motivated assumption that @xmath103 .
the scenario considered can be , therefore , ruled out based on the cmb data .
other scenarios , such as those with @xmath104 at the horizon - crossing have to be studied separately , taking into account the evolution of modes in the intermediate energy range .
the results presented in this essay neither prove nor disprove causal dynamical triangulations .
what has been essentially shown is that phenomenology of cdt can be extracted from numerical studies of the spectral dimension .
the predicted energy - dependence of the speed of light allowed us to rule out a possibility of low - energy dimensional reduction .
furthermore , by incorporating the predictions regarding the uv behavior of the spectral dimension to primordial cosmology , we found that the cdt - motivated cosmological scenarios may be observationally falsifiable .
moreover , we explicitly ruled out one of them with use of the observations of the cmb . as thomas a. edison once said `
i have not failed .
i ve just found 10000 ways that wo nt work . ' quoting the sentence , we are convinced that unabated stubbornness in eliminating planck scale scenarios will eventually lead to a discovery comparable only with edison s one .
jm is supported by the grant dec-2014/13/d
/ st2/01895 of the national centre of science .
the author would like to thank to daniel coumbe for his valuable comments .
j. ambjorn , j. jurkiewicz and r. loll , phys .
* 85 * ( 2000 ) 924 .
j. ambjorn , a. goerlich , j. jurkiewicz and r. loll , phys .
* 519 * ( 2012 ) 127 .
j. ambjorn , j. jurkiewicz and r. loll , phys . rev
. lett .
* 93 * ( 2004 ) 131301 .
j. ambjorn , a. grlich , j. jurkiewicz and r. loll , phys .
* 100 * ( 2008 ) 091304 .
j. ambjorn , s. jordan , j. jurkiewicz and r. loll , phys .
d * 85 * ( 2012 ) 124044 .
j. ambjorn , j. jurkiewicz and r. loll , phys .
lett . *
95 * ( 2005 ) 171301 .
d. n. coumbe and j. jurkiewicz , arxiv:1411.7712 [ hep - th ] .
s. carlip , aip conf .
* 1196 * ( 2009 ) 72 . g. amelino - camelia , l. freidel , j. kowalski - glikman and l. smolin , phys .
d * 84 * ( 2011 ) 084010 . t. p.
sotiriou , m. visser and s. weinfurtner , phys .
d * 84 * ( 2011 ) 104018 .
g. amelino - camelia , j. r. ellis , n. e. mavromatos , d. v. nanopoulos , s. sarkar and , nature * 393 * ( 1998 ) 763 .
v. vasileiou , a. jacholkowska , f. piron , j. bolmont , c. couturier , j. granot , f. w. stecker and j. cohen - tanugi _ et al .
_ , phys .
d * 87 * ( 2013 ) 12 , 122001 .
g. amelino - camelia , m. arzano , g. gubitosi and j. magueijo , phys .
d * 87 * ( 2013 ) 12 , 123532 .
p. a. r. ade _ et al .
_ [ bicep2 and planck collaborations ] , [ arxiv:1502.00612 [ astro-ph.co ] ] .
a. r. ade _ et al .
_ [ planck collaboration ] , astron .
astrophys .
* 571 * ( 2014 ) a16 . | this essay discusses phenomenological aspects of the diffusion time dependence of the spectral dimension predicted by the causal dynamical triangulations ( cdt ) approach to quantum gravity .
the deformed form of the dispersion relation for the fields defined on the cdt space - time is reconstructed . using the _ fermi _ satellite observations of the grb 090510 source we find that the energy scale of the dimensional reduction is @xmath0 gev at ( 95 @xmath1 cl ) . by applying the deformed dispersion relation to the cosmological perturbations
it is shown that , for a scenario when the primordial perturbations are formed in the uv region , the scalar power spectrum @xmath2 where @xmath3 . here
, @xmath4 is obtained from the cdt value of the spectral dimension in the uv limit and @xmath5 is the tensor - to - scalar ratio .
we find that , the predicted deviation from the scale - invariance ( @xmath6 ) is in contradiction with the up to date _ planck _ and _ bicep2 _ results . |
formation and evolution processes of molecular hydrogen ( @xmath0 ) and interstellar dust in galaxies can be strongly coupled , because the surface of dust grains can be the major formation sites of @xmath0 ( e.g. , gould & salpeter 1963 ; hollenbach & salpeter 1971 ) .
dust has long been considered to play decisive roles in several aspects of star and galaxy formation , such as radiative cooling processes in star - forming clouds ( e.g. , herbst 2001 ) and the formation of metal - poor low - mass stars in the early universe ( e.g. , schneider & omukai 2010 ) .
likewise , @xmath0 is an essential element in giant molecular clouds where star formation is ongoing ( e.g. , blitz et al .
2007 ; fukui & kawamura 2010 ) and its physical properties ( e.g. , mass densities ) are key parameters for the observed star - formation laws in galaxies ( e.g. , bigiel et al .
2008 ; leroy et al .
thus , the better understanding of the formation and evolution processes of _ both _ dust and @xmath0 in ism can lead to the deeper understanding of galaxy formation and evolution in general . physical properties of @xmath0 have long been investigated for galaxies with different hubble types ( e.g. , young & scoville 1991 , ys91 ; boselli et al . 2014 ) , in different environments ( e.g. , leon et al . 1998 ; wilson et al .
2009 ) , and at different redshifts ( e.g. , daddi eta l. 2010 ; tacconi et al .
2010 ; bauermeister et al .
2013 ) , and their origins have not been clarified yet .
recent extensive observational studies on h i and @xmath0 properties of galaxies and their correlations of galaxy parameters for a large number of galaxy samples have provided valuable information on scaling relations of gas and stars and thus new constraints on galaxy formation and evolution .
( e.g. , catinella et al .
2010 ; saintonge et al .
2011 ) . resolved structures and kinematics of giant molecular clouds ( gmcs ) in nearby galaxies such as m31 , m33 , and the large magellanic cloud ( lmc ) have enabled astronomers to reveal the physical factors for the conversion from h i to @xmath0 and the possible typical lifetime of gmcs in galaxies ( e.g. , blitz et al .
2007 ; fukui & kawamura 2010 ) . in spite of these observational progresses , a number of key long - standing problems on @xmath0 properties of galaxy have not been resolved yet . among them
is the origin of the diverse @xmath0 properties along the hubble sequence ( e.g. , ys91 ) .
it is well known that the mass - ratio of @xmath0 to h i ( referred to as @xmath7 ) is quite diverse ( more than two orders of magnitudes ) for a given hubble type and it is systematically higher in early - type disk galaxies ( ys91 ) .
this diverse @xmath0 properties along the hubble sequence has been confirmed in the latest observations with a larger number of galaxy samples ( e.g. , boselli et al .
it is theoretically unclear what global galaxy parameters ( e.g. , bulge - to - disk - ratios ) can determine the observed @xmath0-to - h i - ratios ( @xmath7 ) in galaxies .
furthermore , one of other unresolved problems is related to the observed radial distributions of @xmath0 in galaxies and their correlations with stellar parameters of galaxies ( e.g. , blitz et al .
wong & blitz ( 2002 ) showed that the ratio of @xmath0 to h i surface density in a galaxy depends on the projected distance ( @xmath10 ) from the galactic center such that it can be best fit to @xmath8 ( or @xmath11 , see their fig .
some early - type spirals ( e.g. , m31 ) are observed to have intriguing ring - like structures ( e.g. , ys91 ) .
recent observations have revealed that the mass - ratio of @xmath0 to stars ( @xmath12 ) is quite different between galaxies with different hubble types and even within a same type ( e.g. , boselli et al .
these observations have not been explained clearly by previous theoretical studies of galaxy formation and evolution , though they would have some profound implications on galaxy formation .
recent significant progresses in observational studies of @xmath0 properties of galaxies with different types at different @xmath13 appear to have triggered extensive theoretical and numerical studies of of @xmath0 formation in galaxies .
for example , recent numerical simulations of galaxy formation and evolution have incorporated the conversion of @xmath14 formation from h i on dust grains ( e.g. , pelupessy et al .
2006 , p06 ; robertson & kravtsov 2008 , gnedin et al . 2009 ; christensen et al . 2012
; kuhlen et al .
2012 ; thompson et al .
recent semi - analytic models of galaxy formation ( e.g. , fu et al . 2010 ; lagos et al . 2012 )
adopted the @xmath14 formation model dependent gaseous metallicities and densities proposed by krumholz , mckee & tumlinson ( 2009 ) and thereby investigated the time evolution of @xmath14 contents in galaxies .
however , these previous numerical studies of galaxy formation and evolution with @xmath0 formation have not incorporated the evolution of dust ( e.g. , abundances and masses ) explicitly in their models , and accordingly could not discuss the importance of the _ joint _ evolution of dust and @xmath0 .
it is ideal for any theoretical studies of @xmath0 formation in galaxies to include the formation and evolution processes of both dust and @xmath0 in a self - consistent manner , because their evolution can be strongly coupled through star formation .
given that recent observational studies have revealed a number of important correlations between dust and @xmath0 properties in galaxies ( e.g. , corbelli et al .
2012 ) , it would be essential for theoretical studies of @xmath0 formation in galaxies to incorporate a model for the formation and evolution of dust in ism .
our previous studies constructed a new chemodynamical model that includes both the @xmath0 formation on dust grains and the formation and destruction of dust in ism in a self - consistent manner ( bekki 2013a , 2014 ; b13a and b14 , respectively , yozin & bekki 2014a ) .
they therefore could discuss the time evolution of dust and @xmath0 properties in galaxies at their formation epochs .
however , they did not discuss how the physical properties of @xmath0 in galaxies can possibly depend on their global parameters such as their total masses and hubble types .
therefore , the following three questions are unresolved : ( i ) what can drive the observed diversity in @xmath0 properties between different galaxies , ( ii ) whether and how dust can play a role in controlling @xmath0 properties of galaxies , and ( iii ) what is responsible for the observed @xmath0-dust correlations in galaxies .
the purpose of this paper are two - fold as follows .
first , we investigate whether the adopted model of @xmath0 formation self - consistently including dust evolution can explain the observed fundamental properties of @xmath0 in galaxies . in this first investigation
, we focus particularly on the observed correlations between mass - ratios of @xmath0 to h i and other galaxy parameters ( e.g. , wong & blitz 2002 ; blitz & rosolowsy 2006 , blitz et al . 2007 ) .
second , we investigate how the physical properties of @xmath0 in galaxies depend on the total halo masses ( @xmath2 ) , baryonic mass fractions ( @xmath3 ) , gas mass fractions ( @xmath4 ) , hubble types ( e.g. , bulge - to - disk - ratios , @xmath15 ) , and the formation redshifts of galaxies by using mainly isolated models of disk and dwarf galaxies .
we also try to understand how galaxy interaction can influence the time evolution of global @xmath0 properties of galaxies in this second investigation .
the plan of the paper is as follows .
we describe some details of the chemodynamical model with the formation and evolution of dust and @xmath14 adopted by the present study in 2 .
we present the numerical results on the physical correlations between @xmath0 properties and galaxy parameters ( e.g. , masses ) in disk galaxies in 3 . in this section
, we also discuss the dependences of the results on the adopted model parameters . in 4 , we discuss the latest observational results on @xmath14 properties of galaxies and provides some implications of some key results derived in this paper .
we summarize our conclusions in 5 . in the present
, we do not discuss the latest results from @xmath16 ( atacama large millimeter array ) , such as the molecular outflow driven by agn ( e.g. , combes et al . 2014 ) and total ism masses probed by dust emission in 107 galaxies from @xmath17 to @xmath18 ( scoville et al .
2014 ) . these new observational results will be addressed by our future works with a more sophisticated chemodynamical model with @xmath0 formation .
[ cols="<,^ , < " , ]
we adopt our new simulation code recently developed in our previous works ( b13a and b14 ) in order to investigate spatial and temporal variations of @xmath0 in disk galaxies with different masses and hubble - types .
the adopted code can be run on gpu - based machines ( clusters ) , where gravitational calculations can be done on gpus whereas other calculations ( e.g. , star formation and hydrodynamics ) can be done on cpus .
the code adopts the smoothed - particle hydrodynamics ( sph ) method for following the time evolution of gas dynamics in galaxies .
the present simulations include the formation of dust grains in the stellar winds of supernovae ( sne ) and asymptotic giant branch ( agb ) stars , the time evolution of interstellar radiation field ( isrf ) , the growth and destruction processes of dust in the interstellar medium ( ism ) , the @xmath0 formation on dust grains , and the @xmath0 photo - dissociation due to far ultra - violet ( fuv ) light in a self - consistent manner .
although we can investigate the formation of polycyclic aromatic hydrocarbon ( pah ) dust in carbon - rich agb stars and the time evolution of dust by using the simulation code , we do not extensively investigate the dust properties in the present study .
instead , we briefly discuss the possible physical correlations between dust and @xmath0 properties in disk galaxies .
the new simulation code does not include the effects of feedback of active galactic nuclei ( agn ) on ism and the growth of supermassive black holes ( smbhs ) in galaxies so that we can not investigate how the feedback can change the spatial distributions and mass budgets of dust and @xmath0 in galaxies .
although it is an important issue whether and how the agn feedback effects can influence the dust properties of ism and thus the @xmath0 contents , we will discuss this in our future papers .
the new code allows us to choose whether some physical effects ( e.g. , dust formation etc ) are included or excluded ( i.e. , switched on or off ) so that we can investigate how the physical effects are important in the evolution of gas and stars in galaxies .
accordingly , we summarize the physical effects that are included in the present study in table 1 for clarity .
since the major formation site of @xmath0 in ism is the surface of dust grains , it is crucial for any theoretical study on @xmath0 evolution of galaxies to properly investigate the time evolution of dust abundances in disk galaxies . given that both observational and theoretical studies have shown that dust - to - metal - ratios can be diverse ( e.g. , hirashita 1999 ; galametz et al .
2011 ) , we need to refrain from deriving dust abundances from metallicities by assuming constant dust - to - metal - ratios .
the present code indeed enables us to investigate the time evolution of different dust components in galaxies so that we can more properly predict the time evolution of @xmath0 in galaxies .
we mainly focus on isolated disk galaxies composed of dark matter halo , stellar disk , stellar bulge , and gaseous disk : some new results on @xmath0 properties in forming galaxies at high redshifts ( @xmath13 ) are discussed in b14 . since
the details of the physical models for the formation and evolution of dust and @xmath0 are given in b13 and b14 , we briefly describe the models in the present paper . the total masses of dark matter halo , stellar disk , gas disk , and bulge of a disk galaxy are denoted as @xmath2 , @xmath19 , @xmath20 , and @xmath21 , respectively .
the total disk mass ( gas + stars ) and gas mass fraction are denoted as @xmath22 and @xmath4 , respectively , for convenience . the mass ratio of the disk ( @xmath23 ) to the dark matter halo ( @xmath2 ) in a disk galaxy is a ` baryonic mass fraction ' and denoted as @xmath3 .
the bulge - to - disk - ratio is defined as @xmath24 and represented by a parameter @xmath15 .
the four key parameters in the present study are @xmath2 , @xmath4 , @xmath3 , and @xmath15 . in order to describe the initial density profile of dark matter halo in a disk galaxy
, we adopt the density distribution of the nfw halo ( navarro , frenk & white 1996 ) suggested from cdm simulations : @xmath25 where @xmath26 , @xmath27 , and @xmath28 are the spherical radius , the characteristic density of a dark halo , and the scale length of the halo , respectively .
the @xmath29-parameter ( @xmath30 , where @xmath31 is the virial radius of a dark matter halo ) and @xmath31 are chosen appropriately for a given dark halo mass ( @xmath32 ) by using the @xmath33 relation for @xmath34 and 2 predicted by recent cosmological simulations ( e.g. , neto et al . 2007 ; mu@xmath35oz - cuartas et al .
2011 ) .
the bulge of a disk galaxy has a size of @xmath36 and a scale - length of @xmath37 and is represented by the hernquist density profile .
the bulge is assumed to have isotropic velocity dispersion and the radial velocity dispersion is given according to the jeans equation for a spherical system .
the bulge - to - disk ratio ( @xmath38 ) of a disk galaxy is a free parameter ranging from 0 ( pure disk galaxy ) to 4 ( bulge - dominated ) .
the ` mw - type ' models are those with @xmath39 and @xmath40 , where @xmath41 is the stellar disk size of a galaxy .
we adopt the mass - size scaling relation of @xmath42 for bulges so that we can determine @xmath36 for a given @xmath21 .
the value of @xmath43 is determined so that @xmath36 can be 3.5 kpc for @xmath44 ( corresponding to the mass and size of the mw s bulge ) .
the radial ( @xmath10 ) and vertical ( @xmath45 ) density profiles of the stellar disk are assumed to be proportional to @xmath46 with scale length @xmath47 and to @xmath48 with scale length @xmath49 , respectively .
the gas disk with a size @xmath50 has the radial and vertical scale lengths of @xmath51 and @xmath52 , respectively . in the present model for the mw - type , the exponential disk has @xmath53 kpc and @xmath54 kpc .
in addition to the rotational velocity caused by the gravitational field of disk , bulge , and dark halo components , the initial radial and azimuthal velocity dispersions are assigned to the disc component according to the epicyclic theory with toomre s parameter @xmath55 = 1.5 .
the vertical velocity dispersion at a given radius is set to be 0.5 times as large as the radial velocity dispersion at that point .
the total numbers of particles used for dark matter halo ( @xmath56 ) , stellar disk ( @xmath57 ) , and gaseous disk ( @xmath58 ) in a simulation are 700000 , 200000 , and 100000 respectively .
the total number of particle for bulge is @xmath59 , which means that @xmath60 for the mw - type disk galaxy model with @xmath61 and @xmath62 for the big bulge model with @xmath63 .
therefore , the total number of particles is @xmath64 for the fiducial mw - type model and @xmath65 for the big bulge model .
the gravitational softening length for each component is determined by the number of particle used for each component and by the size of the distribution ( e.g. , @xmath41 and @xmath31 ) and the value is described later .
observations have shown that ( i ) there is a mass - metallicity relation in disk galaxies ( e.g. , tremonti et al .
2004 ) and ( ii ) most of disk galaxies show negative metallicity gradients ( i.e. , higher metallicity in inner regions ) with the slopes being different in different galaxies ( e.g. , zaritsky et al .
we therefore determine the initial mean metallicity of the gas disk in a galaxy according to the observe mass - metallicity relation for the adopted total halo mass ( @xmath2 ) and gas mass fraction for the galaxy .
for example , a mw - type disk galaxy with @xmath66 ( @xmath67 ) and @xmath68 ( corresponding to the fiducial model later described ) can have [ fe / h]=0 .
the gas - phase metallicity of each ( gaseous and stellar ) particle is given according to its initial position : at @xmath26 = @xmath10 , where @xmath26 ( @xmath10 ) is the projected distance ( in units of kpc ) from the center of the disk , the metallicity of the star is given as : @xmath69}_{\rm r = r } = { \rm [ fe / h]}_{\rm d , r=0 } + \alpha \times { \rm r}. \,\ ] ] where @xmath70 is the slope of the metallicity gradient in units of dex kpc@xmath71 . since the present results on global properties of @xmath0 do not depend on @xmath70 , we mainly show the results of the models with @xmath72 .
we however discuss briefly how the present results can change if we adopt a steeper metallicity gradient of @xmath73 , which is the observed value of our milky way ( e.g. , andrievsky et al .
2004 ) .
since we mainly investigate @xmath0 properties of disk galaxies at @xmath34 , we construct the disk model by using the @xmath33 relation for dark matter halos at @xmath34 ( e.g. , neto et al . 2007 ;
mu@xmath35oz - cuartas et al .
we however investigate a number of disk models that can represent disk galaxies at high @xmath13 by using the @xmath33 relation dependent on @xmath13 ( mu@xmath35oz - cuartas et al .
2011 ) . in constructing the disk models at high @xmath13
, we consider the following two theoretical and observational results .
first , the mean density of a dark matter halo ( @xmath74 ) with @xmath2 at @xmath13 is @xmath75 times higher than that ( of a dark matter halo with the same @xmath2 ) at @xmath34 ( @xmath76 ) .
second , the stellar disk size of a galaxy at @xmath13 ( @xmath41 ) is @xmath77 times larger than that at @xmath34 ( @xmath78 )
. therefore the mean density of a dark matter halo in a disk galaxy at @xmath34 is as follows ; @xmath79 where @xmath80 for a given @xmath2 .
the stellar disk size is given as follows : @xmath81 which means that the size ratio of @xmath31 to @xmath41 is proportional to @xmath82 .
for example , @xmath29 and @xmath31 for a disk galaxy with @xmath66 at @xmath34 are 10 and 245kpc , respectively , whereas they are 4.5 and 82 kpc at @xmath83 .
the simulated dark matter halos are assumed to have a scaling relation of @xmath84 at @xmath34 and 2 .
since sf can proceed in molecular clouds , we adopt the following ` @xmath0-dependent ' sf recipe ( b13a ) using molecular gas fraction ( @xmath1 ) defined for each gas particle in the present study .
a gas particle _ can be _ converted into a new star if ( i ) the local dynamical time scale is shorter than the sound crossing time scale ( mimicking the jeans instability ) , ( ii ) the local velocity field is identified as being consistent with gravitationally collapsing ( i.e. , div * v*@xmath85 ) , and ( iii ) the local density exceeds a threshold density for star formation ( @xmath86 ) .
we mainly investigate the models with @xmath87 @xmath88 in the present study .
a gas particle can be regarded as a ` sf candidate ' gas particle if the above three sf conditions ( i)-(iii ) are satisfied .
it could be possible to convert some fraction ( @xmath89 ) of a sf candidate gas particle into a new star at each time step until the mass of the gas particle becomes very small .
however , this sf conversion method can increase dramatically the total number of stellar particles , which becomes numerically very costly .
we therefore adopt the following sf conversion method .
a sf candidate @xmath90-th gas particle is regarded as having a sf probability ( @xmath91 ) ; @xmath92 where @xmath93 corresponds to a star formation efficiency ( sfe ) in molecular cores and is set to be 1 , @xmath94 is the time step width for the gas particle , @xmath95 is the gas density of the particle , and @xmath96 is the power - law slope of the kennicutt - schmidt law ( sfr@xmath97 ; kennicutt 1998 ) .
a reasonable value of @xmath98 is adopted in the present study .
this sf probability has been already introduced in our early chemodynamical simulations of galaxies ( e.g. , bekki & shioya 1998 ) . at each time step random numbers ( @xmath99 ; @xmath100 ) are generated and compared with @xmath91 .
if @xmath101 , then the gas particle can be converted into a new stellar one . in this sf recipe , a gas particle with a higher gas density and thus a shorter sf timescale ( @xmath102 ) can be more rapidly converted into a new star owing to the larger @xmath91 .
equally , a gas particle with a higher @xmath1 can be more rapidly converted into a new star .
we thus consider that the present sf model is a good approximation for star formation in molecular gas of disk galaxies . each sn is assumed to eject the feedback energy ( @xmath103 ) of @xmath104 erg and 90% and 10% of @xmath103 are used for the increase of thermal energy ( ` thermal feedback ' ) and random motion ( ` kinematic feedback ' ) , respectively .
the thermal energy is used for the ` adiabatic expansion phase ' , where each sn can remain adiabatic for a timescale of @xmath105 .
although @xmath106 yr is reasonable for a single sn explosion , we adopt a much longer @xmath105 of @xmath107 yr .
this is mainly because multiple sn explosions can occur for a gas particle with a mass of @xmath108 in these galaxy - scale simulations , and @xmath105 can be different for multiple sn explosions in a small local region owing to complicated interaction between gaseous ejecta from different sne .
such interaction of multiple sn explosions would make the adiabatic phase significantly longer in real ism of galaxies .
we adopt a canonical stellar initial mass function ( imf ) proposed by kroupa ( 2001 ) , which has three different slopes at different mass ranges .
our recent study ( bekki 2013b , b13b ) has shown that if the imf depends on local physical properties of ism ( e.g. , gas density and metallicity ) , then galaxy evolution , in particular , the time evolution of @xmath0 and dust contents , can be significantly influenced by the time - varying imf . since the main purpose of this study is not to demonstrate how the @xmath0 contents of disk galaxies can be influenced by a time - varying imf , we assume that the imf is fixed at the adopted kroupa - imf in the present study .
it is our future works to clarify the roles of time - varying imfs in controlling @xmath0 properties in galaxies with different masses and hubble types .
since the present model for chemical enrichment processes of galaxies is exactly the same as that used in b13a , we briefly describe the model here .
chemical enrichment through star formation and metal ejection from snia , ii , and agb stars is self - consistently included in the chemodynamical simulations .
we investigate the time evolution of the 11 chemical elements of h , he , c , n , o , fe , mg , ca , si , s , and ba in order to predict both chemical abundances and dust properties in the present study . we consider the time delay between the epoch of star formation and those of supernova explosions and commencement of agb phases ( i.e. , non - instantaneous recycling of chemical elements ) .
we adopt the ` prompt sn ia ' model in which the delay time distribution ( dtd ) of sne ia is consistent with recent observational results by extensive sn ia surveys ( e.g. , mannucci et al . 2006 ) . in this prompt sn ia mode , there is a time delay ( @xmath109 ) between the star formation and the metal ejection for sne ia and the range of @xmath109 is 0.1 gyr @xmath110 10 gyr . the fraction of the stars that eventually produce sne ia for 38@xmath111 has not been observationally determined . in the present study , @xmath112 is assumed .
we adopt the nucleosynthesis yields of sne ii and ia from tsujimoto et al .
( 1995 ; t95 ) and agb stars from van den hoek & groenewegen ( 1997 ; vg97 ) in order to estimate chemical yields in the present study .
since the dust model adopted in the present study is the same as those in b13a and b14 , we here briefly describe the model .
we calculate the total mass of @xmath113th component ( @xmath113=c , o , mg , si , s , ca , and fe ) of dust from @xmath114th type of stars ( @xmath114 = i , ii , and agb for sne ia , sne ii , and agb stars , respectively ) based on the methods described in b13a that is similar to those adopted in dwek ( 1998 , d98 ) .
we consider that the key parameter in dust accretion is the dust accretion timescale ( @xmath115 ) . in the present study
, this parameter can vary between different gas particles and is thus represented by @xmath116 for @xmath90th gas particle .
the mass of @xmath113th component ( @xmath113=c , o , mg , si , s , ca , and fe ) of dust for @xmath90th gas particle at time @xmath117 ( @xmath118 ) can increase owing to dust accretion processes .
the mass increase is described as @xmath119 where @xmath120 is the individual time step width for the @xmath90th gas particle and @xmath121 is the fraction of the @xmath113th chemical element that is locked up in the dust .
owing to this dust growth , the mass of @xmath113th chemical component that is _ not _ locked up in the dust ( @xmath122 ) can decrease , which is simply given as @xmath123 as is clear in these equations , the total mass of @xmath113th component in @xmath90th gas particle ( @xmath124 ) is @xmath125 .
dust grains can be destroyed though supernova blast waves in the ism of galaxies ( e.g. , mckee 1989 ) and the destruction process is parameterized by the destruction time scale ( @xmath126 ) in previous one - zone models ( e.g. , lisenfeld & ferrara 1998 ; hirashita 1999 ) .
following the previous models , the decrease of the mass of @xmath113th component of dust for @xmath90th gas particle at time @xmath117 due to dust destruction process is as follows @xmath127 where @xmath128 is the dust destruction timescale for @xmath90th particle .
the dust destroyed by supernova explosions can be returned back to the ism , and therefore the mass of @xmath113th chemical component that is not locked up in the dust increases as follows : @xmath129 thus the equation for the time evolution of @xmath113th component of metals for @xmath90th gas particle are given as @xmath130 likewise , the equation for dust evolution is given as @xmath131 dust is locked up in stars as metals are done so , when gas particles are converted into new stars .
we investigate the variable dust accretion ( vda ) model in which @xmath132 is different between different particles with different gaseous properties and changes with time according to the changes of gaseous properties .
we need to introduce a few additional parameters in vda in order to describe the possible dependences of @xmath115 of gas particles on the gas densities , temperature , and chemical abundances .
our previous simulations with vda ( b14 ) clearly show the importance of dust accretion and destruction in the evolution of @xmath133 and @xmath1 .
the details of the vda model is given in ( b14 ) , which discusses the comparison between the vda and the constant dust accretion model in which @xmath115 is constant for all particles throughout simulations .
we adopt the following dependence of @xmath116 on the mass density and temperature of a gas particle in the vda : @xmath134 where @xmath135 and @xmath136 are the gas density and temperature of a @xmath90-th gas particle , respectively , @xmath137 ( typical ism density at the solar neighborhood ) and @xmath138 ( temperature of cold gas ) are set to be 1 atom @xmath88 and 20k , respectively , and @xmath139 is a reference dust accretion timescale at @xmath140 and @xmath141 .
the dust destruction timescale , @xmath128 , for each gas particle is described as follows : @xmath142 where @xmath143 controls the timescale ratio of dust destruction to dust growth . in our previous studies ( b13a , b14 ) ,
@xmath144 is demonstrated to reproduce the observed dust properties of galaxies .
we therefore consider that @xmath145 should be adopted a reference value in the present dust model . in the present study ,
the dust - to - metal - ratio of a galaxy is defined as follows : @xmath146 where @xmath147 and @xmath148 are the total amount of dust and metals in the galaxy , respectively .
although the initial value of @xmath149 could be different in different galaxies , we mainly adopt the standard value of 0.4 ( corresponding to the value of the solar neighborhood ) and thereby investigate the evolution of dust and @xmath0 .
we briefly discuss how the present results depend on the initial @xmath149 by changing @xmath149 from 0.1 to 0.4 .
the model for @xmath0 formation and dissociation in the present study is exactly the same as those used in b13a : @xmath0 formation on dust grains and @xmath14 dissociation by fuv radiation are both self - consistently included in chemodynamical simulations . the temperature ( @xmath150 ) , hydrogen density ( @xmath151 ) , dust - to - gas ratio ( @xmath133 ) of a gas particle and the strength of the fuv radiation field ( @xmath152 ) around the gas particle are calculated at each time step so that the fraction of molecular hydrogen ( @xmath1 ) for the gas particle can be derived based on the @xmath0 formation / destruction equilibrium conditions .
thus the @xmath0 fraction for @xmath90-th gas particle ( @xmath153 ) is given as ; @xmath154 where @xmath155 means a function for @xmath153 determination .
since the detail of the derivation methods of @xmath156 and @xmath153 ( thus @xmath155 ) are given in b13a and b13b , we here briefly describe the methods .
the seds of stellar particles around each @xmath90-th gas particles ( thus isrf ) are first estimated from ages and metallicities of the stars by using stellar population synthesis codes for a given imf ( e.g. , bruzual & charlot 2003 ) .
then the strength of the fuv - part of the isrf is estimated from the seds so that @xmath156 can be derived for the @xmath90-th gas particle .
based on @xmath156 , @xmath157 , and @xmath158 of the gas particle , we can derive @xmath153 ( see fig . 1 in b13a ) .
thus each gas particle has @xmath153 , metallicity ( [ fe / h ] ) , and gas density , all of which are used for estimating the imf slopes for the particle ( when it is converted into a new star ) .
the ages of stars ( @xmath159 ) are assumed to be 5 gyr and 2 gyr for the models with @xmath34 and @xmath83 , respectively .
the results of the present models are influenced by the choice of @xmath159 only for the very early - phase of disk evolution ( @xmath160 gyr ) , if @xmath159 is rather short ( @xmath161 gyr ) .
the influences of @xmath159 is very limited , because the photodissociation of @xmath0 by stars can be very efficient when gas is surrounded by very young stars and thus irradiated by the strong uv - radiation fields .
we therefore discuss only the results of the models with @xmath159 adopted above .
a number of previous theoretical models of @xmath0 formation ( e.g. , hidaka & sofue 2002 ) adopted the phase transition theory proposed by elmegreen ( 1993 ) , in which h i - to-@xmath0 transition in ism is determined basically by gas pressure and radiation field .
elmegreen ( 1993 ) already showed that the observed @xmath0 mass function ( @xmath162 ) can be understood by his theory and also suggested that spiral density waves can convert h i into @xmath0 .
we do not adopt his phase transition theory but instead try to reproduce the observed @xmath0 properties using a different model based on @xmath0 formation on dust grains .
one of key ingredients of the code is that the gravitational softening length ( @xmath163 ) is chosen for each component in a galaxy ( i.e. , multiple gravitational softening lengths ) .
thus the gravitational softening length ( @xmath163 ) is different between dark matter ( @xmath164 ) and gas ( @xmath165 ) and @xmath164 is determined by the initial mean separation of dark matter particles .
the gravitational softening length for stars ( @xmath166 ) is set to be the same as that for gas .
initial @xmath165 is set to be significantly smaller than @xmath164 owing to rather high number - density of gas particles . the softening length for new stars formed from gas
is set to be the same as @xmath165 .
furthermore , when two different components interact gravitationally , the mean softening length for the two components is applied for the gravitational calculation .
for example , @xmath167 is used for gravitational interaction between the values of @xmath164 and @xmath165 ( @xmath168 ) are 2.1 kpc and 0.2 kpc , respectively , for the fiducial mw - type disk model .
we consider that the ism in galaxies can be modeled as an ideal gas with the ratio of specific heats ( @xmath169 ) being 5/3 . the gaseous temperature ( @xmath150 ) is set to be @xmath170 k initially in all models .
the basic methods to implement sph in the present study are essentially the same as those proposed by hernquist & katz ( 1989 ) .
we adopt the predictor - corrector algorithm ( that is accurate to second order in time and space ) in order to integrate the equations describing the time evolution of a system .
each particle is allocated an individual time step width ( @xmath94 ) that is determined by physical properties of the particle . the maximum time step width ( @xmath171 )
is @xmath172 in simulation units , which means that @xmath173 yr in the present study .
although a gas particle is allowed to have a minimum time step width of @xmath174 yr in the adopted individual time step scheme , no particle actually has such a short time step width ( @xmath175 yr ) owing to conversion from gas to star in high - density gas regions .
the radiative cooling processes are properly included by using the cooling curve by rosen & bregman ( 1995 ) for @xmath176k and the mapping iii code for @xmath177k ( sutherland & dopita 1993 ) . in order to investigate how external tidal perturbation can influence the physical properties of @xmath0 in disk galaxies , we investigate tidal interaction models in which two disk galaxies strongly interact with each other .
one of the two galaxies ( ` primary galaxy ' ) is represented by the disk galaxy model described above whereas the interacting companion galaxy is represented by a point - mass particle .
although the mass - ratio of the companion to the primary can be a free parameter represented by @xmath178 , we present the results only for the models with @xmath179 in which the influences of tidal interaction on @xmath0 properties can be clearly seen . in all of the simulations of tidal interaction ,
the orbit of the two disks is set to be initially in the @xmath180 plane and the distance between the center of mass of the two disks is set to b @xmath181 ( corresponding to 175 kpc for the mw - type disk model ) for most models .
the pericenter distance , represented by @xmath182 , is set to be @xmath183 .
the eccentricity is set to be 1.0 for all models of galaxy interaction , meaning that the encounter of galaxy interaction is parabolic .
the spin of the primary galaxy in an interacting pair is specified by two angle @xmath184 and @xmath185 ( in units of degrees ) , where @xmath184 is the angle between the @xmath13 axis and the vector of the angular momentum of a disk and @xmath185 is the azimuthal angle measured from @xmath186 axis to the projection of the angular momentum vector of a disk on to @xmath180 plane . in the present study ,
we present the results of the following four tidal interaction models : ( i ) prograde ( ` pr ' ) model with @xmath184 = 0 , @xmath185 = @xmath187 , and @xmath188 , ( ii ) retrograde ( ` re ' ) model with @xmath184 = 180 , @xmath185 = 0 , and @xmath188 ( iii ) highly inclined model with @xmath184 = 30 , @xmath185 = @xmath189 , and @xmath188 , and ( iv ) distant interaction model with @xmath184 = 0 , @xmath185 = @xmath187 , and @xmath190 .
we investigate numerous models to try to answer the following key questions related to the origin of @xmath0 in disk galaxies : ( i ) how can @xmath0 gas distribute within disks ? , ( ii ) what determines the molecular gas ( @xmath0 ) fractions in disk galaxies ? , ( iii ) what are the roles of galactic bulges in controlling @xmath0 properties of disk galaxies ? , ( iv ) can dust evolution influence the evolution of h i and @xmath0 in galaxies ? , ( v ) does galaxy interaction enhance @xmath0 contents of galaxies ? , ( vi ) is there any threshold gas density ( or threshold galaxy mass ) beyond which @xmath0 formation is possible ? , and ( vii ) can high-@xmath13 disk galaxies can contain larger fractions of @xmath0 within disks ? it should be noted here that these are selected , because the present numerical code allows us to investigate these : other questions such as the influences of agn and dust size evolution on @xmath0 evolution can not be addressed in the present study . in order to address these questions
, we mainly describe the results of the selected 34 models in the present study .
we indeed investigated more than 34 models , in particular , for low - mass disk models in order to understand whether there can be a threshold halo mass beyond which @xmath0 formation is possible .
however , we discuss only these results , because these representative models can more clearly show how the results depend on model parameters .
the values of parameters adopted in these models are given in table 2 . in order to discuss the importance of dust evolution in @xmath0 evolution of galaxies , we investigate the models with different basic dust parameters .
the parameter values for dust models and chemical evolution are summarized in table 3 .
we first describe the results of the ` fiducial model ' ( m1 ) that corresponds to a disk galaxy similar to the milky way ( thus referred to as mw - type disk model ) .
then we discuss how the key model parameters controls the @xmath0 properties of disk galaxies in detail .
the present chemodynamical models do not have enough spatial resolution to resolve the sub - pc - scale central cores of gmcs with @xmath0 .
we therefore can not discuss whether and where self - gravitating molecular clouds are formed in galactic gas disks . instead of investigating the molecular cloud mass function ( mcmf )
, we try to derive the @xmath0 surface density distribution ( h2sdd ) , which would be useful in discussing the origin of mcmfs of galaxies .
we estimate the h2sdd in a disk galaxy as follows .
we first divide the simulated stellar disk region ( @xmath53 kpc for the mw model ) of a disk galaxy into @xmath191 or @xmath192 small meshes and thereby estimate the total mass of @xmath0 ( @xmath193 ) .
we count the number of meshes ( @xmath194 ) with @xmath193 within a given mass range so that we can estimate the h2sdd .
the mesh size can be as small as @xmath195pc for the mw - type disk models , which can be enough to discuss the mass of each individual @xmath0 gas clouds in a local region of a disk galaxy .
although this method is not exactly the same as that used in the observational estimation of mcmf , we consider that the derived h2sdd could be quite useful in discussing the origin of the observed mcmf in a qualitative manner ( e.g. , discussing the slope of the mcmf ) .
we mainly check whether the simulated slope of h2sdd can be similar to the observed slope of the galactic mcmf .
figs . 1 and 2 describe how the spatial distributions of stars , h
i , @xmath0 , and dust evolve with time during the dynamical evolution of the gas disk with spiral arms and a central bar in the fiducial mw - type disk model ( m1 ) .
numerous spiral arms can be developed at @xmath196 gyr owing to gravitational instability , and they can influence the conversion from h i to @xmath0 in the gas disk after their formation .
the locations of high - density @xmath0 regions at @xmath196 , 0.8 , and 1.1 gyr appear to be roughly coincident with the locations of h i spiral arms of the gas disk .
however , @xmath0 shows a very clumpy distribution along spirals arms , which is in a striking contrast with the relatively smooth distribution of h i along spiral arms .
this implies that only the high - density parts of the gaseous spiral arms can be the possible formation sites of @xmath0 in galactic disks .
numerous @xmath0 clumps can be active star - forming regions in the present sf model based on @xmath0 gas densities .
although the projected distribution of interstellar dust follows the spiral - like distribution of h i gas , it does not show clump - like structures as @xmath0 .
interstellar dust can become less clumpy , because it can be efficiently destroyed by star - formation ( i.e. , sne ) which can occur in high - density gas clumps .
the physical roles of spiral arms in the formation of gmcs have been already discussed in a number of recent theoretical works based on high - resolution numerical simulations of gas disk evolution in disk galaxies ( e.g. , dobbs et al .
2011 ; wada et al . 2011
; halle & combes 2013 ) .
although these previous works did not self - consistently model the @xmath0 formation on dust grains and the time evolution of dust , they have already shown projected @xmath0 distributions in galactic disks ( e.g. , fig .
14 in halle & combes 2013 ) .
the derived distributions of @xmath0 along spiral arms in the present mw - type disk model are similar to those reported in halle & combes ( 2013 ) .
the clumpy @xmath0 distributions derived in the present study are not so clearly seen in previous works , which could reflect the differences in the models of sf and sn feedback effects between the present and other works .
3 shows the spatial distributions of h i and @xmath0 and the mass function of @xmath0 gas clouds ( referred to as molecular cloud mass function ; mcmf ) in the gas disk at @xmath197 gyr in the fiducial mw - type disk galaxy model with @xmath68 .
the simulated @xmath0 distribution is more centrally concentrated than h i gas within the disk owing to the more efficient formation of @xmath0 on dust grains in the inner part of the disk where both gas density and @xmath133 can be higher during the disk evolution .
the spatial distribution of @xmath0 gas shows small - scale clumpy structures within gaseous spiral arms where gas density can become rather large .
clearly , there is an outer truncation ( @xmath198 kpc ) beyond which no / little @xmath0 gas can be found in this model .
the projected radial density profile of @xmath0 is steeper than that of h i in this model .
the central part of the disk is dominated by @xmath0 gas and most of the gas particles there have rather high @xmath199 ( @xmath200 ) , which implies that molecular hydrogen can be systematically more massive .
the simulated h2sdd has a slope of @xmath201 for @xmath202 @xmath203 pc@xmath204 , which means that the simulated slope is similar to the observed slope of mcmf that is approximated as @xmath205 for the mw and m31 ( e.g. , fig . 4 in fukui & kawamura 2010 ) .
this implies that the present model of @xmath0 formation on dust grains is realistic and reasonable for @xmath0 formation in disk galaxies .
although the mcmf can be shifted toward higher masses if a coarser mesh size ( @xmath206 ) is used for the mcmf derivation , the slope of the mcmf does not change significantly .
this mw - type fiducial model can develop a central stellar bar owing to bar instability and the stellar bar can influence the formation process of @xmath0 significantly .
the formation process of the stellar bar in the central 2 kpc and its influence on gas for @xmath207 gyr are given in appendix a. fig .
4 shows that the gas disk has a very strong concentration of @xmath0 in the central 200 pc at @xmath208 gyr .
this @xmath0 concentration corresponds to the formation of nuclear @xmath0 gas disk due to the dynamical action of the central stellar bar on the surrounding gas . owing to the rather high concentration of @xmath0 in the disk ,
the sfr can be significantly increased , in particular , in the central region of this barred disk galaxy .
furthermore , the @xmath0 fraction can be significantly increased ( from @xmath209 to @xmath210 ) .
thus these results clearly demonstrate that stellar bars can play a vital role in converting h i to @xmath0 in disk galaxies .
5 describes how the simulated h2sdds depend on the model parameters for metallicity gradient and dust growth / destruction .
the slopes of the h2sdds do not depend strongly on the parameters , though some models ( with steep metallicity gradient and longer dust accretion / destruction timescales ) show the overproduction of high - density @xmath0 regions with @xmath211 @xmath203 pc@xmath204 within their disks .
the larger number of high - density @xmath0 regions is due largely to the stronger central concentration of @xmath0 gas in the inner regions of the gas disks .
as shown in fig .
4 , a central region of a galaxy is dominated by high - density @xmath0 gas so that almost all of the meshes in the central region can have high @xmath212 . as a result of this
, the simulated h2sdds can have a bump around @xmath213 @xmath203 pc@xmath204 .
it is a bit surprising that the simulated h2sdds are less sensitive to the adopted dust model .
6 describes the physical correlations between molecular fraction ( @xmath7 ) , which is defined as the mass ratio of @xmath0 to h i , and @xmath214 , where @xmath215 is the gas pressure of a particle and @xmath216 is the boltzmann constant , in the gas disk of the fiducial mw model with different initial metallicity gradients @xmath72 and @xmath217 dex kpc@xmath71 at @xmath218 gyr .
it should be stressed that this @xmath7 is not exactly the same as @xmath1 ( @xmath219 ) , which is often shown in figures of this paper .
the observed correlation of @xmath9 ( e.g. , blitz et al .
2007 ; fukui & kawamura 2010 ) can be used as a key observational constraint for the present @xmath0 formation model .
although the dispersion in @xmath7 at each bin appears to be large , the two models show a clear positive correlation between @xmath7 and @xmath215 ( i.e. , higher @xmath7 for higher gaseous pressure ) .
this result implies that the present model for @xmath0 formation , which does not explicitly assume a dependence of @xmath0 formation efficiency on gaseous pressure , does a good job in predicting @xmath7 and its dependence on gaseous pressure .
although the two models with different slopes of gaseous metallicity gradients ( @xmath72 and @xmath217 ) have the slopes of the @xmath220 that are similar to the observed one ( @xmath221 ) , the model with a steep negative gradient of gaseous metallicity ( @xmath73 ) can have the slope more similar to the observed one .
the model with @xmath73 shows a lower @xmath0 formation efficiency at @xmath222 ( @xmath88 k ) in comparison with the model with @xmath72 , mainly because the model with a negative metallicity gradient has a significantly lower dust - to - gas - ratio ( i.e. , lower metallicity ) at the outer region of the disk with lower gaseous pressure so that @xmath0 formation efficiency on dust grains can become lower .
the derived weaker dependence of the slope of the @xmath220 relation on initial metallicity gradients implies that the slopes can be quite similar between different galaxies with different metallicity gradients .
7 describes the simulated @xmath220 relations in models with different model parameters for dust growth and destruction .
these models have the parameter values being less realistic ( e.g. , rather long dust accretion timescale ) in comparison with the models used in fig .
6 so that we can discuss how the dust modeling is important for reproducing the observed @xmath220 relation .
it is clear from fig . 7
that the simulated @xmath220 relation in the model with shorter dust growth timescale ( @xmath223 gyr ) is less consistent with the observed one owing to rather small @xmath7 at higher @xmath215 ( @xmath224 ) in the model . the model with a shorter dust destruction timescale ( @xmath225 and @xmath226 gyr )
can not reproduce the observed @xmath220 relation so well either .
these results imply that both dust accretion and destruction timescales should be carefully chosen for reproducing the observed @xmath220 relation .
the models with longer dust accretion timescale ( @xmath227 and 2 gyr ) show @xmath220 similar to the observed one as long as @xmath144 is adopted . however these models
can not reproduce the observed @xmath220 better than the models in fig .
6 , which implies that not only @xmath145 but also @xmath228 can be a key parameter for reproducing the observed @xmath220 . the slope of the simulated @xmath220 in the model with a longer dust destruction time scale ( @xmath229 and @xmath226 gyr )
can better match the observed one except for rather high gaseous pressure ( @xmath230 ) .
the model with an initially lower dust - to - metal - ratio ( @xmath231 ) shows @xmath220 relation similar to those in fig .
6 , though @xmath7 at higher @xmath215 is systematically smaller than those in fig .
6 owing to the lower @xmath0 formation efficiency caused by the lower dust - to - gas ratio .
the models with no dust evolution show significantly shallower @xmath220 relations , irrespective of the slopes of the initial radial metallicity gradients ( @xmath72 and @xmath217 ) .
this result implies that the time evolution of dust abundances in disk galaxies needs to be included for the self - consistent reproduction of the observed @xmath220 relation .
the models without sn feedback ( snf ) and those without chemical evolution show significant deviation from the observed @xmath220 , which strongly suggests that dust growth and destruction processes caused by chemical enrichment and sn explosions need to be carefully included in reproducing the observed @xmath220 relation in a quantitative manner . thus these results demonstrate that as long as the @xmath0 formation processes on dust grains are included in hydrodynamical simulations of galaxy formation and evolution , the dust growth and destruction processes , which can be influenced by chemical enrichment and sn explosions , should be carefully considered in discussing the origin of the observed @xmath220 relation .
the observed radial gradients of @xmath7 in disk galaxies ( e.g. , wong & blitz 2002 ) can be an additional constraint for any model for @xmath0 formation in galaxy - scale simulations .
we here estimate @xmath7 at each radius in a simulated disk galaxy by calculating the ratio of @xmath212 ( i.e. , the projected surface mass density of @xmath0 ) to @xmath232 at each radius in fig .
although individual gas particles corresponding to individual local gaseous regions in a disk galaxy have vastly different @xmath7 ( ranging from 0 to 10 ) , the slopes of the radial @xmath7 profiles in the two models with different @xmath70 appear to be similar to the observed one ( @xmath233 ) . the model with @xmath72 can slightly better reproduce the observed slope than the model with @xmath73 .
thus , the simulated slopes similar to the observed @xmath233 demonstrate that the present model for @xmath0 formation on dust grains is quite reasonable and realistic in discussing the global ( galaxy - scale ) distributions of neutral and molecular hydrogen in disk galaxies ( as long as a reasonable set of model parameters for dust is adopted ) .
a number of recent observational studies have found intriguing correlations between physical properties of dust , gas , and stars in galaxies ( e.g. , corbelli et al 2012 ; cortese et al .
2012 ; boselli et al .
it is therefore essential for the present study to provide some comparisons between these observations and the corresponding simulation results .
shows that the @xmath0 fractions ( @xmath1 ) are more likely to be larger for local regions with higher stellar surface densities ( @xmath234 ) , in particular , for @xmath235 kpc@xmath204 .
there is a large dispersion in @xmath1 for a given @xmath234 , which reflects the fact that star formation histories and chemical enrichment processes ( thus dust growth / destruction processes ) are quite different in different local regions . the local regions with higher @xmath234 can have higher gas densities so that conversion from h i to @xmath0 can occur more efficiently
this is a physical reason for the simulated trend of @xmath1 with @xmath234 in fig .
there is a positive correlation between the mass - ratio of @xmath0 to stars and @xmath234 , though the correlation can not be simply described as a power - law profile owing to the flatter slope in the simulated relation at @xmath235 kpc@xmath204 .
a large dispersion in @xmath12 can be seen , which again reflects the diverse histories of star formation and chemical enrichment in different local regions .
the simulated relation between @xmath212 and @xmath236 is relatively tight in comparison with other two correlations and can be described roughly as @xmath237 .
these simulated relations between physical properties of dust , gas , and stars are at least qualitatively consistent with observational results ( e.g. , boselli et al .
. the physical properties of @xmath0 such as radial profiles of @xmath7 and time evolution of @xmath1 can be quite diverse in the simulated disk galaxies with different model parameters .
since it is not so meaningful ( and much less productive ) to describe the results of all models in detail , we here describe only the results that are quite important and thus worth mentioning . in the following subsections , @xmath0 properties ( e.g. , @xmath1 )
are compared between models with different values of a model parameter ( e.g. , @xmath4 ) for a given set of other model parameters ( i.e. , other model parameters are fixed ) .
although the morphologies of @xmath0 distributions in disk galaxies depend largely on the adopted @xmath15 and @xmath4 in disk galaxies , almost all models in the present study shows very strong central concentration of @xmath0 and clear coincidence between the locations of gas particles with high @xmath0 fractions and those of spiral arms in the disks .
the distributions of @xmath0 along gaseous spiral arms appear to be very clumpy , which reflects the fact that only the high - density parts of the spiral arms can efficiently form @xmath0 .
these clumpy distributions are in a striking contrast with the h i distributions in disk galaxies .
furthermore , the detailed spatial distributions of @xmath0 in the central few kpc of the disks can be controlled by the presence ( or the absence ) of stellar bars developed from bar instability in the central regions .
10 shows one of clear examples of the strong influences of the central stellar bars on the spatial distributions of @xmath0 in disk galaxies . clearly , the disk galaxy in this model ( m8 ) has a very elongated ring - like structure dominated by @xmath0 , which is formed as a result of dynamical action of the central bar on the gas disk .
this kind of ` molecular gas ring ' can be seen in the models in which @xmath15 is not so large and thus stellar bar formation can not be severely suppressed by the central big bulges .
the time evolution of @xmath1 depend strongly on @xmath4 , @xmath3 , @xmath15 , @xmath2 , and the formation redshifts of galaxies in the present models .
furthermore , tidal galaxy interaction can significantly change @xmath238 and @xmath1 depending on the orbit configurations of interacting two galaxies . the time evolution of 2d @xmath232 and @xmath212 distributions for interacting galaxies is discussed in appendix b , because the formation process of @xmath0 is quite interesting for interacting galaxies .
the final 2d @xmath232 and @xmath212 distributions for a high-@xmath13 disk model are in a striking contrast with those of low-@xmath13 disks and thus shown in appendix b. since the time evolution of @xmath238 and @xmath1 are very similar in each model , we briefly summarize the derived key 7 dependences of @xmath1 ( not @xmath238 ) on the model parameters in figs .
11 and 12 .
firstly , @xmath1 can be initially larger in the models with larger @xmath4 for a given set of other model parameters ( e.g. , @xmath15 ) .
the final values of these two quantities ( @xmath218 gyr ) in the model with larger @xmath4 ( = 0.27 ) , however , can become rather small owing to rapid @xmath0 gas consumption by star formation . as a result of this
, the model with larger @xmath4 can have smaller final @xmath238 and @xmath1 than the model with smaller @xmath4 ( = 0.09 ) .
it should be noted that the final @xmath4 in the model with @xmath239 can become rather low too .
secondly , @xmath1 can be initially smaller in the models with smaller @xmath3 for a given @xmath4 ( and @xmath2 ) , mainly because formation of non - axisymmetric structures such as bars and spirals can be suppressed in less strongly self - gravitating disks ( owing to the dominant dark matter halos ) so that @xmath0 formation efficiency can be lower .
the lower surface gas densities in the models with smaller @xmath3 are also responsible for the smaller @xmath1 .
the smaller @xmath1 for smaller @xmath3 can be seen in low - mass models with @xmath240 .
as shown in the model with smaller @xmath3 ( 0.02 ) yet larger @xmath4 ( = 0.33 ) , @xmath1 can be larger than the model with @xmath241 and smaller @xmath4 . since gas consumption by star formation is slower in these models with smaller @xmath3 , the initial @xmath1 difference in the models ( i.e. , larger @xmath1 for larger @xmath4 ) can last for a long timescale .
thirdly , @xmath1 can be _ initially _ larger in the models with smaller bulges , because bigger galactic bulges can slow down or suppress the formation of spiral arms and bars so that @xmath0 formation efficiencies in gas disks can be also severely suppressed by the bulges .
12 shows that although @xmath1 is initially larger in the model with @xmath39 ( fiducial model with a smaller bulge ) than in the model with @xmath242 ( @xmath243 gyr ) , the @xmath1 difference becomes smaller as the time passes by . after the formation of a stronger bar in the central region of the fiducial model , the bar dynamically acts on the gas disk so that a larger amount of gas can be transferred to the central region , where @xmath0 formation is enhanced .
consequently , @xmath1 can become significantly larger at @xmath244 gyr .
this kind of @xmath1 enhancement by the dynamical action of a bar can occur to a lesser extend in the big bulge model with @xmath242 in which only a weak stellar bar can be formed .
it should be stressed here that the models ( m9 and 12 ) with rather large @xmath15 ( 2 and 4 ) , initial @xmath1 is rather low ( @xmath245 ) .
however , the @xmath0 consumption by star formation is much less efficient in the models so that @xmath1 can keep slowly rising ( owing to the higher gas densities ) . as a result of this
, these models can finally show larger @xmath1 ( @xmath246 ) at @xmath197 gyr .
therefore , we can not claim that disks with bigger bulges show smaller @xmath1 in the present study .
fourthly , low surface brightness ( lsb ) disk galaxies can have initially smaller @xmath238 and @xmath1 owing to their low gas densities .
it should be stressed that owing to much less rapid gas consumption and chemical enrichment , the lsb model can finally have @xmath238 and @xmath1 similar to those of the fiducial model .
fifthly , galaxy interaction between two disk galaxies can significantly increase @xmath1 after the pericenter passage .
this enhancement of @xmath0 formation in disks can be more clearly seen in the ` prograde ' tidal interaction model ( pr ) in which the spin axes of two disks are roughly parallel to those of their orbital angular momentum .
this enhancement of @xmath0 formation , however , can not be clearly seen in other interaction models ( re or distant encounter models ) , which implies that not all of interacting galaxies show significantly large @xmath1 in comparison with isolated field disk galaxies .
galaxy interaction can also change the distribution of @xmath0 and the mcmf , in particular , for the central region .
the mcmfs of interaction models are discussed in appendix a. sixthly , less massive disk galaxies can have smaller @xmath1 , firstly because the formation of bars and spirals arms can be severely suppressed in the galaxies ( owing to the stronger concentration of dark matter and smaller @xmath3 ) , and secondly because the dust - to - gas ratios ( @xmath133 ) , which are key factors for @xmath0 formation models of the present study , can be lower in the disks . the chemical enrichment processes by sne and agb stars in these low - mass models are so slow ( owing to low sfrs ) that @xmath133 can not increase during the dynamical evolution of the disks .
therefore , the formation efficiencies of @xmath1 can be kept lower in these models .
the low - mass model with @xmath240 and smaller @xmath247 shows rather small @xmath1 , which means that this low - mass disk galaxy is almost @xmath0-less . the baryonic fraction ( @xmath3 )
is particularly important in determining the time evolution of @xmath1 for low - mass disk galaxies .
seventhly , disk galaxies at higher @xmath13 can have larger @xmath1 than those at @xmath34 , mainly because disks are more compact so that a larger amount of @xmath0 gas can be formed in the high - density gaseous regions of the disks .
it should be noted that ( i ) the simulated mw - type disk with low @xmath68 at @xmath83 can have a larger @xmath1 ( @xmath248 ) and ( ii ) even the less massive disk with @xmath249 can have a larger @xmath1 ( @xmath250 ) .
these results imply that high-@xmath13 disk galaxies are highly likely to have systematically rather high @xmath0 gas fractions .
the dependences of @xmath238 and @xmath1 on @xmath4 and @xmath2 can be seen among the simulated disk galaxies at @xmath83 .
13 shows that the five mw - type disk models with different @xmath4 , @xmath3 , @xmath15 , and surface stellar densities have qualitatively similar projected radial profiles of h i , @xmath0 , and @xmath7 , though the absolute magnitudes of these surface mass densities are different between the models .
the @xmath212 profiles are steeper than @xmath232 in the five models , and the radial profiles of @xmath7 ( @xmath251 ) have negative slopes ( i.e. , higher @xmath7 at smaller @xmath10 ) .
the mw - type disk models with different @xmath15 have the steeper slopes approximated by @xmath8 whereas the model with higher @xmath4 ( = 0.55 ) and smaller @xmath241 and the lsb model have shallower @xmath7 profiles . the lower @xmath212 in the model with larger @xmath4 is due largely to rapid @xmath0 consumption in the inner disk . fig .
13 furthermore shows that tidal galaxy interaction can increase both @xmath232 and @xmath212 significantly owing to the strong gaseous concentration formed in the central regions of the interacting disks and the tidal arms .
the strong central @xmath0 concentration is due largely to efficient gas - transfer to the nuclear region induced by tidal bar and gaseous dissipation .
the formation of @xmath0 is more efficient in the central region and
tidal arms during galaxy interaction , which is discussed in more detail in appendix b. the radial profile of @xmath7 in the interaction model at @xmath252 kpc is rather flat and there is the lack of @xmath0 gas in the circumnuclear region ( @xmath253 kpc ) in the model .
the less massive disk models with @xmath240 and @xmath254 have shallower @xmath212 profiles thus shallower @xmath7 profiles too , which is in a striking contrast with the mw - type disk models .
the origin of the low @xmath7 in low - mass disk models is closely associated with weak / no spiral and bar structures ( due to the less strongly gravitation disks ) and initially low @xmath133 .
14 clearly demonstrates that if both @xmath4 and @xmath3 are the same between the fiducial mw - type disk model and the low - mass dwarf disk model ( m25 with @xmath240 ) , then @xmath212 is dramatically different between the two models .
the maximum @xmath255 is only @xmath256 @xmath203 pc@xmath204 in the low - mass disk model , which means that even if baryonic fractions are rather high , gas - poor low - mass disk galaxies are unlikely to form @xmath0 efficiently .
a more gas - rich dwarf disk model shows a higher @xmath212 along s - shaped gaseous region , but @xmath212 is much lower than the mw - type disk model .
the high - z disk models ( @xmath83 ) also have shallower @xmath7 profiles in the present study .
as shown in preceding sections , a strong central concentration of @xmath0 can occur in the barred disk galaxies owing to the dynamical action of stellar bars on gas .
after the rapid consumption of @xmath0 gas in the centrally concentrated gas in the barred disk galaxies , there can be a deficiency of @xmath0 in their central regions .
it is confirmed that the bulge - less disk galaxy with @xmath257 at @xmath258 gyr ( i.e. , after long - term evolution ) , for which a strong bar can be spontaneously formed owing to global bar instability , can finally show the lack of @xmath0 in its central region after rapid @xmath0 consumption by star formation in the central region .
this result implies that if gas accretion onto disks is truncated , even the late - type disk galaxies with no / little bulges can show the lack of @xmath0 gas in their central regions .
15 describes the physical correlations of local properties of @xmath0 with those of dust and stars within galaxies with different model parameters .
these ` internal ' @xmath0-scaling relations will be able to be compared with ongoing observational studies on @xmath0 properties of disk galaxies ( by , e.g. , alma ) , and thus the predicted scaling relations will be useful for interpreting the observational results .
it is clear that the models with different @xmath4 , @xmath15 , and initial stellar mass densities of the disks have a similar positive correlation between @xmath12 and @xmath234 for @xmath259 @xmath203 kpc@xmath204 .
the simulated relatively flat correlation for @xmath260 @xmath203 kpc@xmath204 is due to the strong concentration of @xmath0 in the central regions ( where @xmath234 is high ) caused by dynamical action of the stellar bars on gas in the disks .
the tidal interaction model , in which @xmath1 can be enhanced and @xmath0 distribution can be significantly changed , shows a @xmath261 relation similar to those seen in the isolated mw - type disk models , though the correlation is not so clear owing to the disturbed @xmath0 distribution .
the less massive disk models with low @xmath2 also show strong positive @xmath261 correlations for lower @xmath234 , though the central @xmath0 concentration makes the correlation shallower for higher @xmath234 .
the high-@xmath13 models with @xmath83 also have clear positive @xmath261 correlations , which implies that a linear @xmath261 correlation will be identified in future observations of @xmath0 distributions in high@xmath262 disk galaxies .
these ten models show a much tighter correlation between @xmath212 and @xmath236 , which means that local ism with higher dust surface densities can have higher @xmath0 surface densities in disk galaxies .
this simulated almost linear @xmath263 scaling relation is not so surprising , because @xmath0 formation efficiencies in local ism are assumed to depend on the dust - to - gas ratios of local ism in the present study .
recent observational studies have just started a detailed investigation on the sub - kpc - scale properties of dust in nearby disk galaxies like m31 ( e.g. , viaene et al . 2014 ) , which can be combined with the observed sub - kpc - scale properties of @xmath0 to produce a possible dust-@xmath0 scaling relation within galaxies . the observed @xmath263 scaling relation within galaxies
will be able to be used as a stringent test for any theoretical model of @xmath0 formation on dust grains in galaxies .
16 describes the dependence of the maximum values of @xmath238 and @xmath1 during @xmath264 gyr dynamical evolution on the initial halo masses ( @xmath2 ) and dust - to - gas - ratios ( @xmath133 ) in the selected models .
it is clear from this figure that the models with larger @xmath2 or larger @xmath133 are more likely to have larger @xmath265 and @xmath266 the low - mass models have initially smaller @xmath133 due to the adopted mass - metallicity relation , and spiral arms and bars , which are the major driver for efficient @xmath0 formation in disks , can not be formed owing to the less strongly self - gravitating disks ( i.e. , more strongly dominated by dark matter ) .
this combination of the low @xmath133 and the incapability of bar / spiral formation in the low - mass models is the main reason for the derived dependences of @xmath265 and @xmath266 on @xmath2 and @xmath133 .
it should be stressed here that some models with @xmath267 and those with @xmath268 have no @xmath0 in the disks and thus no star formation in the present star formation models in which new stars can form exclusively from @xmath0 gas .
this means that low - mass disk galaxies with very low @xmath133 are dominated by h i gas and have no star formation so that they can be identified as isolated massive h i clouds with no star formation ( in other words , dark h i galaxies ) .
the total @xmath0 mass can not become larger than @xmath269 in some of low - mass models with @xmath270 , which means that massive star clusters and globular clusters are unlikely to be formed from gmcs in low - mass disk galaxies .
we will discuss this point later in the discussion section .
one of the unresolved problems related to @xmath0 contents of galaxies dependent on hubble types is that sc / sd late - type disks have smaller @xmath1 ( @xmath7 ) in spite of their having larger gas masses with respect to their dynamical masses ( @xmath271 , see figs . 4 and 5 in ys91 , as an example for this ) .
although this problem will be able to be better discussed in our forthcoming papers in which @xmath272 cdm galaxy formation models with @xmath0 formation are adopted , we can discuss this problem based on the present new results .
galactic bulges have been demonstrated to suppress @xmath0 formation in ism only for the early dynamical evolution of disk galaxies , if @xmath15 is quite large ( @xmath273 , corresponding to sa / s0 galaxies ) .
the main reason for this is that spiral arms and bars , which can enhance the @xmath0 formation , can be suppressed by the presence of big bulges .
the stronger isrfs of big and dense bulges can be also responsible for less efficient conversion from h i to @xmath0 in ism
. such a suppression effect of @xmath0 formation can be seen in models with different @xmath2 and @xmath4 .
however , the larger @xmath1 in disks with smaller bulges can not last long owing to the more rapid gas consumption by star formation .
furthermore , big bulge can severely suppress the star formation so that @xmath1 in disks with big bulges can steadily increase owing to chemical enrichment ( i.e. , increasing @xmath133 ) and finally become significantly larger .
this means that although @xmath1 can be larger in disks with smaller bulges _ for a fixed _ @xmath4 , such disks do not necessarily continue to show larger @xmath1 ( particularly when there is no external gas supply ) .
also the initially larger @xmath1 in disks with smaller bulges appears to be inconsistent with the observed trend of @xmath1 with hubble types .
given these results , it appears difficult to understand the origin of smaller @xmath1 in late - type sc / sd galaxies in the context of the bulge effects on @xmath0 formation .
the central stellar bars formed from global bar instability have been demonstrated to enhance significantly @xmath1 ( @xmath7 ) , in particular , in the central regions of disk galaxies , though such enhancement can not last long owing to rapid gas consumption by star formation .
strong spirals arms have been demonstrated to be where @xmath0 formation is very efficient in the present study . given that global bar / spiral instability can be prohibited by big bulges , disks with later types ( i.e. , smaller bulges ) are more likely to have bars and thus show larger @xmath1 .
this possible trend of @xmath1 with hubble types is opposite to the observed trend , which implies that the formation efficiency of stellar bars is not the major factor that determines @xmath1 in sc / sd galaxies .
the present study has also shown that the final @xmath1 of disk galaxies depend on the initial baryonic mass fractions ( @xmath3 ) such that @xmath1 can be larger in the models with larger @xmath3 for a given @xmath4 ( @xmath2 ) .
recent observations have shown that @xmath3 can be smaller in galaxies with lower halo masses ( e.g. , papastergis et al .
these observations combined with the present results therefore imply that if the late - type sc / sd galaxies can be preferentially formed from dark halos with lower @xmath2 , then the observed systematically small @xmath1 can be more clearly explained .
although the observed luminosity functions for different hubble types in the local field and the virgo cluster suggests that sa+sb galaxies are systematically brighter than sc+sd galaxies ( e.g. , binggeli et al . 1988 ) , observational studies have not yet determined the mass functions of dark matter halos and baryonic fractions for different hubble types .
therefore , it is unclear whether the above explanation for the observed small @xmath1 in sc / sd galaxies is really plausible and realistic , though the smaller @xmath3 in later hubble types would be a promising scenario .
the present study has shown that low - mass disk galaxies show rather small @xmath274 ( maximum possible @xmath0 fraction ) and therefore can be identified as h i - dominated galaxies .
the major physical reason for this very low @xmath0 contents is that these low - mass galaxies are extremely dust - poor so that they can not form @xmath0 on dust grains .
the possible smaller @xmath3 in these low - mass galaxies can also contribute to the very low @xmath0 formation efficiency of the ism .
furthermore , some of the simulated low - mass galaxies with rather high initial @xmath4 ( @xmath6 ) can show virtually @xmath275 , if @xmath2 is as low as or lower than @xmath276 .
these results imply that there would exist low - mass , and extremely low surface - brightness galaxies with only h i gas : these are almost ` dark ' galaxies .
so far no observational studies have found strong evidence for the presence of _ isolated _ ( or ` intergalactic ' ) dark h i - rich galaxies with total gas masses larger than @xmath269 in the nearby universe , though a number of observational studies investigated h i contents of galaxies in group environments ( e.g. , kilborn et al 2005 ; pisano et al .
although most of the detected h i objects have optical counterparts ( or tidal origin ) in these observations , a few observations so far discovered massive apparently isolated h i gas clouds with masses larger than @xmath277 ( e.g. , davies et al . 2004 ; kolibalski et al . 2004 ) .
these intriguing massive isolated objects are likely to be just tidal debris rather than dark h i galaxies ( e.g. , bekki et al . 2005 ) .
if h i - dominated low - mass dark galaxies do not exist , the present study implies that such low - mass galaxies had lost almost all of their gas at their formation epochs or during their evolution owing to some physical processes such as ram pressure stripping and cosmic reionization .
the present study has shown that ( i ) some low - mass galaxies with @xmath267 can not form @xmath0 efficiently owing to their low dust masses and consequently ( ii ) the total @xmath0 masses are unlikely to exceed @xmath278 .
these results can have the following implications on the formation of massive star clusters , in particular , globular clusters ( gcs ) , the origin of which remains unexplained .
as recent theoretical studies of gc formation have demonstrated ( e.g. , dercole et al .
2008 ; bekki 2011 ) , the initial stellar masses of gcs should be @xmath279 times larger than the present ones ( typically , @xmath280 ) : here we do not discuss the physical reasons for the proposed rather high initial masses of gc progenitor .
this means that the original gmcs from which the initial massive stellar systems can form should be larger than @xmath281 even for a high star formation efficiency of @xmath282 .
these theoretical results on gc formation combined with the present results imply that low - mass galaxies with @xmath267 are unlikely to form gcs within them owing to the low @xmath0 contents
. it would be possible that only low - mass , low - density star clusters , can form in their host low - mass galaxies .
these low - mass clusters are prone to tidal destruction by their host galaxies so that they can finally become field stars within their hosts .
if there exists a threshold halo mass above which gcs can form , then it can not only explain the observed absence of gcs in some faint dwarf galaxies in the local group ( e.g. , van den bergh 2000 ) but also provide a physical basis on the minimum halo mass of gc host galaxies introduced in recent semi - analytic models of gc formation in galaxies based on hierarchical galaxy formation scenarios ( e.g. , bekki et al .
2008 ; griffen et al . 2010 ) .
we have investigated the physical properties of @xmath0 in disk galaxies with different masses and hubble types based on our new model for the formation of @xmath0 on dust grains .
the basic parameters in this numerical study are @xmath2 ( halo mass ) , @xmath3 ( baryonic mass fraction ) , @xmath4 ( gas mass fraction ) , and @xmath15 ( bulge - to - disk - ratio ) .
we have investigated how the spatial distributions , time evolution , and scaling relations of @xmath0 depend on these parameters both for isolated disk galaxies and for interacting ones .
the principal results are as follows .
+ ( 1 ) the observed positive correlation between the mass ratio of @xmath0 to h i ( @xmath7 ) and gaseous pressure ( @xmath215 ) can be reproduced reasonably well by the present models . however
, the @xmath220 relation ( @xmath9 ) can be reproduced well only for some ranges of model parameters for dust growth and destruction .
these results imply that dust can play a significant role in the formation of the @xmath220 relation and thus that dust formation and evolution needs to be included in discussing the origin of @xmath0 scaling relations .
+ ( 2 ) the simulated @xmath0 surface density distributions ( h2sdds ) in isolated luminous disk galaxies can have the slope of @xmath201 for @xmath283 @xmath203 pc@xmath204 .
the simulated slope is similar to the slope of the observed molecular cloud mass function for gmcs in the mw ( @xmath284 ) .
although the simulated slopes of h2sdds does not depend so strongly on galaxy parameters such as @xmath15 and @xmath285 in isolated models , the mean @xmath212 can be significantly increased in tidal interaction models .
+ ( 3 ) the simulated radial profiles of the @xmath212-to- @xmath232-ratios in isolated luminous disk galaxies can be described as @xmath8 , which are observed in disk galaxies .
this successful reproduction of observations implies that the present dust - regulated @xmath0 models can grasp some essential ingredients of @xmath0 formation in ism of disk galaxies .
the radial gradients can be different between low - mass and high - mass disk galaxies at low and high @xmath13 and between isolated and interacting galaxies .
the simulated @xmath0 distributions are clumpier than h i in most models
. the @xmath0 surface density of a disk galaxy can dramatically drop beyond a certain radius where @xmath0 formation efficiency becomes rather low owing to the low gas density and the low @xmath133 .
+ ( 4 ) the mass - ratios of @xmath0 to stars ( @xmath12 ) in individual local regions of a galaxy can anti - correlate with the local surface stellar density ( @xmath234 ) .
this anti - correlation ( @xmath286 ) can be seen in different galaxy models with different parameters .
the @xmath0 surface densities ( @xmath212 ) of local individual regions in a galaxy can correlate with the dust surface densities ( @xmath236 ) such that @xmath237 .
this @xmath0-dust correlation is rather tight in different galaxy models .
+ ( 5 ) the initial total masses of dark halos ( @xmath2 ) can control the @xmath0 mass fractions ( @xmath1 ) of disk galaxies in such a way that @xmath1 can be larger for larger @xmath2 .
this is firstly because the formation of spiral arms and bars , where @xmath0 formation is rather efficient , can be severely suppressed in smaller @xmath2 , and secondly because the dust - to - gas - ratios ( @xmath133 ) , which determines the formation efficiency of @xmath0 on dust grains , are lower for smaller @xmath2 .
some low - mass galaxies with @xmath267 show no / little @xmath0 formation within their disks in the present study .
+ ( 6 ) galactic bulges can strongly suppress the formation of @xmath0 in gas disks at the early dynamical evolution of disk galaxies , if @xmath15 is quite large ( @xmath287 ) .
this is mainly because big bulges can prevent global spiral / bar instability from occurring so that efficient @xmath0 formation within spirals and bars can be severely suppressed .
this , however , does not necessarily mean that @xmath1 can continue to be low , because slow gas consumption by star formation yet steady chemical enrichment ( i.e. , increasing @xmath133 ) can finally enhance @xmath1 for some disk models with big bulges in the present study .
these results imply that the origin of some late - type disk galaxies ( sc and sd ) with smaller @xmath1 is not simply due to their small bulges .
some disk models in the present study show the lack of @xmath0 gas in their central regions .
+ ( 7 ) barred disk galaxies are more likely to have larger @xmath1 owing to the formation of centrally concentrated @xmath0 rings and disks .
the central @xmath0 can be rapidly consumed by star formation so that higher @xmath1 in barred galaxies can not last long without further external gas supply onto the disks .
some of the simulated barred disk galaxies can have elongated @xmath0-rings in their central regions , though such @xmath0 structures are short - lived .
bulge - less disk galaxies with central bars can also show the lack of @xmath0 in their central regions after rapid consumption of @xmath0 transferred to the inner few kpc .
+ ( 8) disk galaxies at higher @xmath13 are more likely to have larger @xmath1 owing to their initially high gas densities in the more compact disks .
this result does not depend on other model parameters such as @xmath2 .
disk galaxies with smaller @xmath3 at higher @xmath13 can still show higher @xmath1 . for a given @xmath2 , @xmath4 , and @xmath13 , disk galaxies with smaller @xmath3 have smaller @xmath1 .
high-@xmath13 disks are more likely to have very clumpy @xmath0 distributions with numerous small @xmath0 clouds .
\(9 ) the maximum @xmath0 masses ( @xmath265 ) and fractions ( @xmath266 ) depend on initial halo masses @xmath2 and dust - to - gas - ratios ( @xmath133 ) such that both can be higher in larger @xmath2 and larger @xmath133 . for low - mass halos with @xmath267 , the formation of @xmath0 can be completely suppressed in some models with smaller gas fractions so that no stars can be formed . some low - mass galaxies with @xmath288 can not form @xmath0 owing to rather low dust contents even if they are quite gas - rich @xmath289 .
these galaxies with no @xmath0 are still very rich in h i and could be observationally identified as gas - rich extremely low - surface brightness galaxies ( or almost dark galaxies ) .
+ ( 10 ) the spatial distributions of @xmath0 in the simulated galaxies are quite diverse depending on initial @xmath2 and hubble types ( e.g. , bulge - to - disk - ratio , @xmath15 ) .
circumnuclear @xmath0 rings can be formed in barred disk galaxies , and such ring formation can significantly increase @xmath1 in the galaxies .
radial @xmath0 profiles can be steeper for more luminous disk galaxies in isolation , and tidal interaction can dramatically increase the degrees of central @xmath0 concentration in disk galaxies .
+ ( 11 ) the present results have several important implications of star formation processes of galaxies .
for example , massive star clusters , like old globular clusters , are unlikely to be formed in low - mass galaxies ( @xmath290 ) , because the maximum possible total @xmath0 masses can be well less than @xmath269 ( i.e. , owing to the incapability of giant molecular clouds to form ) .
this result implies that there would exist a threshold galaxy mass above which old gcs can be formed within galaxies at high @xmath13 .
i ( kenji bekki ; kb ) am grateful to the referee for constructive and useful comments that improved this paper .
numerical simulations reported here were carried out on the three gpu clusters , pleiades , fornax , and gstar kindly made available by international center for radio astronomy research ( icrar ) at the university of western australia , ivec , and the center for astrophysics and supercomputing in the swinburne university , respectively .
this research was supported by resources awarded under the astronomy australia ltd s astac scheme on swinburne with support from the australian government .
gstar is funded by swinburne and the australian government s education investment fund .
kb acknowledges the financial support of the australian research council throughout the course of this work .
andrievsky , s. m. , luck , r. e. , martin , p. , lpine , j. r. d. 2004 , a&a , 413 , 159 bauermeister , a. , et al .
2013 , apj , 768 , 132 bekki , k. , 2011 , 412 , 2241 bekki , k. , 2013 , mnras , 432 , 2298 ( b13a ) bekki , k. , 2013 , mnras , 436 , 2254 ( b13b ) bekki , k. , submitted to mnras ( b14 ) bekki , k. , shioya , y. , 1998 , apj , 497 , 108 bekki , k. koribalski , b. s. , & kilborn , v. a. , 2005 , mnras , 363 , l21 bekki , k. , yahagi , h. , nagashima , m. , forbes , duncan a. , 2008 , mnras , 387 , 1131 bigiel , f. , leroy , a. , walter , f. , brinks , e. , de blok , w. j. g. , madore , b. , thornley , m. d. , 2008 , aj , 136 , 2846 binggeli , b. , sandage , a. , tammann , g. a. , 1988 , ara&a , 26 , 509 blitz , l. , rosolowsky , e. , 2006 , apj , 650 , 933 blitz , l. , fukui , y. , kawamura , a. , leroy , a. , mizuno , n. , rosolowsky , e. , 2007 , protostars and planets v , b. reipurth , d. jewitt , and k. keil ( eds . ) , university of arizona press , tucson , 951 pp .
, p.81 - 96 boselli , a. , et al . 2014 , a&a , 564 , 67 bruzual , g. , charlot , s. , 2003 , mnras , 344 , 1000 catinella , b. , et al .
2010 , mnras , 403 , 683 christensen , c. , et al .
, 2012 , mnras , 425 , 3058 combes , f. , et al .
2014 , a&a , 565 , 97 corbelli , e. , 2012 , a&a , 542 , 32 ( c12 ) cortese , l. , et al .
2012 , a&a , 540 , 52 davies , j. , et al .
2004 , mnras , 349 , 922 dercole , a. , vesperini , e. , dantona , f. , mcmillan , s. l. w. , & recchi , s. 2008 , mnras , 391 , 825 daddi , e. , et al .
, 2010 , apj , 713 , 686 dobbs , c. l. , burkert , a. , pringle , j. e. , 2011 , mnras , 417 , 1318 dwek , e. , 1998 , apj , 501 , 643 ( d98 ) elmegreen , b. g. , 1993 , apj , 412 , 90 fukui , y. , kawamura , a. , 2010 , ara&a , 48 , 547 galametz , m. , madden , s. c. , galliano , f. , hony , s. , bendo , g. j. , sauvage , m. , 2011 , a&a , 532 , 56 gould , r. j. , salpeter , e. e. , 1963 , apj , 138 , 393 griffen , b. f. , drinkwater , m. j. , thomas , p. a. , helly , j. c. , pimbblet , k. a. , 2010 , mnras , 405 , 375 halle , a. combes , f. , 2013 , a&a , 559 , 55 herbst , e. , 2001 , chem , soc , rev , 30 , 168 hernquist , l. , & katz , n. , 1989 , 70 , 419 hidaka , m. , sofue , y. , 2002 , pasj , 54 , 223 hirashita , h. , 1999 , apj , 522 , 220 hollenbach , d. , salpeter , e. e. , 1971 , apj , 163 , 155 lagos , c. p. , lacey , c. g. , baugh , c. m. , 2012 , in prepring ( arxiv1210.4974 ) kennicutt , r. c. , jr . , 1998 , apj , 498 , 541 kilborn , v. a. , koribalski , b. s. , forbes , d. a. , barnes , d. g. , musgrave , r. c. , 2005 , mnras , 356 , 77 kroupa , p. , 2001 ,
mnras , 322 , 231 koribalski , b. s. , et al .
2004 , aj , 128 , 16 krumholz , m. r. , mckee , c. f. , tumlinson , j. , 2009 , apj , 693 , 216 leon , s. , combes , f. , menon , t. k. , 1998 , a&a , 330 , 37 leroy , a. k. , et .
al . , 2011 ,
apj , 737 , 12 lisenfeld , u. , ferrara , a. , 1998 , apj , 498 , 145 mckee , c. f. , 1989 , in iau symp . 135 , interstellar dust , edited by louis j. allamandola and a. g. g. m. tielens , p43 [ mannucci et al .
[ ] mannucci , f. , della valle , m. , panagia , n. , 2006 , mnras , 370 , 773 [ mannucci et al .
[ ] mannucci , f. , della valle , m. , panagia , n. , 2006 , mnras , 370 , 773 maoz , d. , sharon , k. , gal - yam , a. , 2010 , apj , 722 , 1879 mori , m. , yoshii , y. , nomoto , k. , 1999 , apj , 511 , 585 mu@xmath35oz - cuartas , j. c. ; macci , a. v. , gottlber , s. , dutton , a. a. , 2011 , mnras , 411 , 584 navarro , j. f. , frenk , c. s. , white , s. d. m. , 1996 , apj , 462 , 563 ( nfw ) neto , a. f. , 2007 , mnras , 381 , 1450 papastergis , e. , cattaneo , a. , huang , s. , giovanelli , r. , haynes , m. p. , 2012 ,
apj , 759 , 138 pelupessy , f. i. , papadopoulos , p. p. , van der werf , p. , 2006 ,
apj , 645 , 1024 ( p06 ) pisano , d. j. , barnes , d. g. , staveley - smith , l. , gibson , b. k. , kilborn , v. a. , freeman , ken c. , 2011 , apjs , 197 ,
28 rosen , a. , bregman , j. n. , 1995 , apj , 440 , 634 saintonge , a. , et al . , 2011 , mnras , 415 , 61 schmidt , m. , 1959 , apj , 129 , 243 schneider , r. , omukai , k. , 2010 , mnras , 402 , 429 scoville , n. , et al .
2014 , apj , 783 , 84 shull , j. m. , 1978 , apj , 226 , 858 stinson , g. , seth , a. , katz , n. , wadsley , j. , governato , f. , quinn , t. , 2006 , mnras , 373 , 1074 sutherland , r. s. , dopita , m. a. , 1993 , apjs , 88 , 253 tacconi , l. j. , et al .
, 2010 , nat , 463 , 781 theis , c. , orlova , n. , 2004 , a&a , 418 , 959 thornton , k. , gaudlitz , m. , janka , h .- th . , steinmetz , m. , 1998 , apj , 500 , 95 tsujimoto , t. , nomoto , k. , yoshii , y. , hashimoto , m. , yanagida , s. , thielemann , f .- k .
, 1995 , mnras , 277 , 945 ( t95 ) van den hoek , l. b. ; groenewegen , m. a. t. , 1997 , a&as , 123 , 305 ( vg97 ) van den bergh , s. , 2000 , the galaxies of the local group viaene , s. , et al .
2014 , accepted in a&a , ( rrxiv:1403.4272 ) wada , k. , baba , j. , saitoh , t. r. , 2011 , apj , 735 , 1 wilson , c. d. , et al .
, 2009 , apj , 693 , 1736 wong , t. , blitz , l. , 2002 , aph , 569 , 157 young , j. s. , scoville , n. z. , 1991 , ara&a , 29 , 581 yozin , c. , bekki , k. , 2014a , accepted in mnras ( arxiv1406.1642 ) yozin , c. , bekki , k. , 2014b , mnras , 439 , 1948 zaritsky , d. , kennicutt , r. c. , jr . , huchra , j. p. , 1994
, apj , 420 , 87
in the present fiducial mw - type disk model ( m1 ) , a stellar bar can play a significant role in the formation of @xmath0 on dust grains and the dynamical evolution of @xmath0 gas , in particular , in the later evolution of the central region of the gas disk .
we here briefly describe the bar formation process in this model . as shown in fig .
a1 on the time evolution of stellar surface density ( @xmath291 ) of the disk , a stellar bar can start to develop around @xmath292 gyr owing to bar instability in this model with a relatively small bulge mass .
the formation of a strong nuclear bar can be completed by @xmath197 gyr so that the bar can influence the gas dynamics there after its formation ( e.g. , a massive nuclear @xmath0 disk can be formed at @xmath258 gyr as a result of dynamical interaction between the bar and @xmath0 gas , as shown in fig .
the gaseous component ( h i+@xmath0 ) also shows a bar - like distribution in the central region of the disk when the stellar bar forms at @xmath197 gyr .
figs . b1 and b2 show intriguing @xmath0 distributions of galactic disks in selected models . in fig .
b1 , the initial @xmath212 of each mesh at @xmath293 is calculated by assuming that @xmath0 mass fraction is 0.01 just for convenience , because @xmath0 formation efficiency based isrf etc is not calculated at @xmath293 .
the formation efficiency of @xmath0 on dust grains can be significantly enhanced in the strong tidal arms of the interacting disk galaxy for the prograde interaction model ( m16 ) , as shown in fig .
b1 ( at @xmath196 gyr ) . during the efficient gas infall onto the nuclear regions , both h
i and @xmath0 gas densities can become significantly higher so that star formation rates can be high in the nuclear region and the tidal arms .
after tidal interaction , the @xmath0 distribution appears to be clumpy with a number of off - center high - density @xmath0 clumps in the disk .
such off - center massive gas clumps can be also seen in recent models of gas - rich interacting galaxies by yozin & bekki ( 2014b ) .
since the number of tidal interaction models investigated in the present study is rather small ( @xmath294 ) , we can not conclude whether such a clumpy @xmath0 distribution is a characteristic of post - interacting galaxies .
we will investigate this point in our future works by performing a larger number of numerical simulations of interacting galaxies with a much wider range of model parameters for galaxy interaction .
b2 shows that a gas - rich , more compact disk at @xmath83 can have a widespread spatial distribution of @xmath0 gas in the model m15 . in this model ,
the baryonic mass fraction is lower than the fiducial mw - disk model ( i.e. , less strongly self - gravitating ) so that the formation of a strong bar can be suppressed .
therefore , a bar - like distribution of gas can not be clearly seen in this model . even in the outer part of the gas disk , high - density @xmath0 regions can be seen , which is in a striking contrast with the @xmath0 distribution of the low-@xmath13 fiducial mw - disk model ( m1 ) .
the gas disk at @xmath83 appears to be composed of numerous small @xmath0 clumps , which implies that the mass function of molecular clouds ( mcmf ) can be significantly different between @xmath34 and @xmath83 .
since the present simulation does not have enough resolution to investigate mcmfs , we will discuss this point in our future numerical works with a much larger number of gas particles . | we investigate the physical properties of molecular hydrogen ( @xmath0 ) in isolated and interacting disk galaxies with different masses and hubble types by using chemodynamical simulations with @xmath0 formation on dust grains and dust growth and destruction in interstellar medium ( ism ) .
we particularly focus on the dependences of @xmath0 gas mass fractions ( @xmath1 ) , spatial distributions of h i and @xmath0 , and local @xmath0-scaling relations on initial halo masses ( @xmath2 ) , baryonic fractions ( @xmath3 ) , gas mass fractions ( @xmath4 ) , and hubble types .
the principal results are as follows .
the final @xmath1 can be larger in disk galaxies with higher @xmath2 , @xmath3 , and @xmath4 .
some low - mass disk models with @xmath2 smaller than @xmath5 show extremely low @xmath1 and thus no / little star formation , even if initial @xmath4 is quite large ( @xmath6 ) .
big galactic bulges can severely suppress the formation of @xmath0 from h i on dust grains whereas strong stellar bars can not only enhance @xmath1 but also be responsible for the formation of @xmath0-dominated central rings .
the projected radial distributions of @xmath0 are significantly more compact than those of h i and the simulated radial profiles of @xmath0-to - h i - ratios ( @xmath7 ) follow roughly @xmath8 in mw - type disk models .
galaxy interaction can significantly increase @xmath1 and total @xmath0 mass in disk galaxies .
the local surface mass densities of @xmath0 can be correlated with those of dust in a galaxy .
the observed correlation between @xmath7 and gas pressure ( @xmath9 ) can be well reproduced in the simulated disk galaxies .
[ firstpage ] ism : molecules galaxies : ism galaxies : evolution infrared : galaxies stars : formation |
the absorption spectra of pure alkali - metal vapors at temperatures of the order 1000 k can be a rich source of information on molecular potentials and properties . achieving a high vapor pressure of lithium in experiments requires higher temperatures than the other alkali - metal atoms , but there are some data from heat pipe ovens @xcite and from a specialized apparatus @xcite .
in addition to the atomic lines the spectra exhibit gross molecular features attributable to transitions between bound levels of the ground electronic state and levels of the excited singlet states and weaker features arising from analogous triplet transitions . theoretically , the envelope of the alkali - metal molecular absorption spectra can be quantitatively reproduced using semi - classical models @xcite and ro - vibrational structure @xcite and continua @xcite can be reproduced from quantum - mechanical models . in this paper
we calculate quantum - mechanically absorption spectra for the @xmath0@xmath1 and @xmath0@xmath2 transitions in @xmath3 .
although both semi - classical @xcite and quantum - mechanical @xcite calculations have been performed and compared @xcite previously for these transitions of @xmath3 , recent improvements in the molecular data prompt the present comprehensive study . from photoassociation spectroscopy and cold collision studies performed in the last few years as well as recent theoretical work
there have been significant critical tests of and improvements to the molecular potentials@xcite , particularly at long - range @xcite , and transition dipole moment data @xcite available for @xmath3 , as well as to the value of the lifetime of the li @xmath4 state @xcite .
this paper presents calculations of the spectra over the full range of wavelengths where absorption in the @xmath0@xmath1 and @xmath0@xmath2 bands is possible .
we calculate the satellite feature profiles at various temperatures , identify and explore the influence of quasibound states and the contributions of bound bound versus bound free transitions , calculate partition functions , and calculate lifetimes for the @xmath1 and @xmath2 ro - vibrational levels .
in the quantum - mechanical formulation an absorption cross section from a vibration - rotation state of the lower electronic state @xmath5 to the vibration - rotation state of the upper electronic state @xmath6 is @xmath7 where @xmath8 is a line - shape function of dimension @xmath9 , @xmath10 is the hnl - london factor and @xmath11 is the transition frequency @xcite . in this study
, @xmath8 is approximated by @xmath12 , with @xmath13 the bin size . for a bound - free transition ,
the absorption cross section from a bound level of the lower electronic state @xmath14 to a continuum level of the upper electronic state @xmath15 can be written as @xmath16 where the continuum wave function @xmath17 is energy normalized .
free - bound or free - free transitions are not considered since the temperatures studied here are not high enough for these types of transitions to be important within the singlet manifold
. the radial wave function can be obtained from the schrdinger equation for the relative motion of the nuclei @xmath18 where @xmath19 is the rotationless potential energy for the electronic state , @xmath20 is the reduced mass of the @xmath21 atoms , and @xmath22 is for bound states the eigenvalue @xmath23 measured with respect to the dissociation limit associated with the wave function @xmath24 .
similarly , for continuum states @xmath22 is the relative kinetic energy of the colliding atoms @xmath25 associated with the energy - normalized wave function @xmath26 .
the total absorption cross section at frequency @xmath27 can be obtained by averaging over initial vibration - rotation levels @xmath5 with a relevant weighting factor and summing over all possible bound - bound and bound - free transitions with frequencies between @xmath27 and @xmath28 @xcite yielding @xmath29 \right .
\nonumber \\ & & \left .
+ \sum_{j'j''v '' } \int d\epsilon ' \sigma_{v''j''\lambda''}^{\epsilon'j'\lambda ' } \omega_{j '' } ( 2j''+1)\exp[-(d_e+e_{v''j''\lambda''})/kt ] \right ] , \end{aligned}\ ] ] where @xmath30 is a statistical factor due to nuclear spin with the values @xmath31=\case{3}{8}$ ] for even @xmath32 and @xmath33=\case{5}{8}$ ] for odd @xmath32 , for @xmath34 with @xmath35 . with the zero of energy taken to be the potential minimum ,
the partition function @xmath36 of the lower state with dissociation energy @xmath37 is @xmath38 , \ ] ] assuming thermodynamic equilibrium .
the resulting cross sections can be used to model the dimer absorption spectra in the quasistatic limit .
the adopted molecular potentials of the @xmath0 , @xmath1 and @xmath2 states are shown in fig . [ potentials ] .
the ground x@xmath39 state potential was constructed using the approach of ref .
we adopted the recommended @xcite potential obtained by barakat _
et al . _
@xcite , who applied the rydberg - klein - rees ( rkr ) method to measured energies .
the data were connected smoothly to the long - range form @xmath40 where @xmath41 is the exchange energy @xcite and the coefficients @xmath42 , @xmath43 , and @xmath44 have been calculated in refs .
@xcite and we use atomic units in this section .
we adopted the coefficients @xmath45 , @xmath46 and @xmath47 from ref .
the two regions were connected at @xmath48 yielding a value of 8516.95 @xmath49 for the dissociation energy of the x@xmath50 state of @xmath51li@xmath52 , in satisfactory agreement with the accepted value of 8516.61 @xmath49 @xcite .
the form @xmath53 was used to extrapolate the potential at short - range where the constants were determined to smoothly connect to the innermost rkr points .
the resulting @xmath0 potential yields an @xmath54-wave scattering length of @xmath55 , in excellent agreement with the accepted @xcite value of @xmath56 and a sensitive test of the assembled data . for the excited a@xmath57 state
the rkr potential of ref .
@xcite was adopted and smoothly connected at about @xmath58 to the long - range form @xmath59 with coefficients @xmath60 and @xmath61 from ref .
@xcite and @xmath62 from ref . @xcite .
for @xmath63 the data were connected to the short range form @xmath53 .
the dissociation energy for the a@xmath64 state is determined to be @xmath65 @xmath49 in our calculation , in good agreement with the experimental value @xcite of @xmath66 @xmath49 given in ref .
@xcite .
the potential for the b@xmath67 state has a hump with maximum at @xmath68 and various determinations of the hump location and height have been summarized in refs .
@xcite and @xcite .
we adopted the ipa ( inverted perturbation approach ) potentials from hessel and vidal @xcite for @xmath69 and the _ ab initio _ potential of schmidt - mink _ et al . _
@xcite for @xmath70 with one additional point at @xmath71 fixing the barrier maximum energy at @xmath72 above dissociation . at @xmath73
the data were connected to the long - range form of eq .
( [ e : along - range ] ) by shifting down the data by @xmath74 from @xmath75 .
the values of the coefficients used in eq .
( [ e : along - range ] ) were @xmath76 and @xmath77 from ref .
@xcite and @xmath78 from ref .
the potential energy data for @xmath69 were fixed using the @xmath2 state dissociation energy of @xmath79 @xmath49 , which we determined using the experimental value for @xmath80 of @xmath81 @xmath49 @xcite , the atomic asymptotic energy of @xmath82 @xmath49 , and the @xmath0 dissociation energy of @xmath83 @xmath49 @xcite . finally , the data in the range @xmath84 were smoothly connected using cubic splines .
for @xmath85 the data were extrapolated using the short range form @xmath53 . transition dipole moments for the @xmath0@xmath1 and @xmath0@xmath2 transitions are available from _ ab initio _ calculations @xcite and for x a transitions from measured lifetimes of a state levels @xcite . for the electronic transition dipole moments
@xmath86 , we adopted _ ab initio _ calculations of ratcliff , fish , and konowalow @xcite connected at @xmath87 to the long - range asymptotic form @xmath88 the value of the coefficient @xmath89 was @xmath90 for both x a and x b transitions and the coefficient @xmath91 was @xmath92 or @xmath93 @xcite for , respectively , x a or x b transitions . for both transitions , we multiplied @xmath94 calculated using the above coefficients by a constant such that the value @xmath95 was identical to the corresponding _ ab initio _ value from ref .
@xcite to provide a smooth connection between short and long - range forms for @xmath86 .
the x a dipole moment function that we adopted is consistent with that derived by baumgartner _
et al . _
@xcite from experimental measurements .
there is no experimentally - derived dipole moment function for the x b transition .
bound and continuum wave functions were calculated using the numerov method to integrate eq .
( [ e : schrodinger ] ) . for the x and a states , eigenvalues were generally in good agreement with the rydberg - klein - rees values used as input to the potentials constructed .
[ there is an apparent misprint for the energy of the @xmath96 level of the x@xmath39 state in ref .
we used @xmath97 @xmath49 , consistent with ref .
the constructed @xmath2 state potential can reproduce the rotationless ipa energies tabulated by hessel and vidal typically to about @xmath98 @xmath49 , with the greatest discrepancy 0.15 @xmath49 for the @xmath99 value .
calculated frequencies of @xmath2@xmath0 transitions were also compared with calculations of verma , koch , and stwalley @xcite and the agreement was good , within 0.1 @xmath49 .
four quasi - bound levels of the @xmath2 state were found .
for the rotationless potential their calculated eigenvalues are 143.91 , 276.7 , 391.74 and 483.88 @xmath49 above the dissociation limit for @xmath100 to @xmath101 .
the calculated term energies of vibration - rotation states in the @xmath0 state were used to compute partition functions using eq .
( [ e : partitionqm ] ) .
the maximum vibrational and rotational quantum numbers in our calculations are 41 and 123 , respectively , for the @xmath0 state . in the harmonic approximation for ro - vibrational energies
, the partition function can be calculated using the simple expression @xmath102\nonumber \\ & & \times \sum_{v''=0}^{41 } \exp[-h\nu_e(v''+1/2)/kt ] , \end{aligned}\ ] ] which assumes that the term energy can be described by the first terms of the power series with respect to vibrational and rotational quantum numbers .
using constants @xmath103 = 351.3904 @xmath49 and @xmath104 @xmath49 @xcite the partition function from eq .
( [ e : partitionlam ] ) was calculated and it is compared with the partition function calculated from eq .
( [ e : partitionqm ] ) for the @xmath0 state as a function of temperature in fig . [ partition ] . the anharmonicity of the potential for higher vibrational levels accounts for the differences between the two results with increasing temperature .
for @xmath105 the @xmath0 state supports quasibound vibrational levels .
the expression eq .
( [ e : partitionqm ] ) for @xmath36 does not specify whether quasibound states are to be included or not in the summations .
we evaluated @xmath36 with and without the quasibound levels to ascertain their importance and the results are shown in fig .
[ partition ] .
the effect of the additional levels becomes increasingly significant with higher temperature .
for the present study covering temperatures between 1000 and 2000 k there is not a significant distinction between @xmath36 , @xmath106 , and the result with the inclusion of the quasibound states .
the molecular fraction can be calculated using the expression @xmath107/[n_{\rm
li}]^2 = ( q_{{\rm li}_2}/q_{\rm li}^2 ) \exp(d_e / kt ) , \ ] ] where the atomic partition function @xmath108 is @xmath109 , with the electronic partition function for the atom well - approximated by the spin degeneracy of 2 for the temperatures studied in the present work , and the molecular partition function @xmath110 is @xmath111 , with the electronic partition function for the @xmath0 state taking the value 1 . the molecular fraction eq .
( [ e : mol - frac ] ) is plotted in fig .
[ partition ] .
the absorption coefficient @xmath112 can be obtained if the atomic density is known from @xmath113 \sigma(\nu ) .\ ] ] lifetimes of the various ro - vibrational levels of the @xmath1 state were measured @xcite and calculated , see for example , refs .
we calculated spontaneous emission transition probabilities and lifetimes of rotational - vibrational levels of the a@xmath57 state in order to test the adopted transition dipole moment .
the spontaneous emission rate from a bound state ( @xmath114 ) to a bound state ( @xmath115 ) is @xmath116 where the electronic state degeneracy is @xmath117 and we have neglected change in the rotational quantum number .
the total spontaneous emission rate from the upper level ( @xmath114 ) can be obtained by summing over all possible transitions to bound and continuum states @xmath118 where @xmath119 is the spontaneous emission probability to a continuum energy @xmath120 with partial wave @xmath121 .
the lifetime is @xmath122 lifetimes of levels @xmath123 of the @xmath1 state were measured by baumgartner _
et al . _
the @xmath1 state is affected by indirect predissociation via the @xmath124 and @xmath125 states @xcite .
the measured lifetimes @xmath126 of vibration - rotation levels thought to be unaffected by indirect predissociation taken from ref .
@xcite and corresponding calculated lifetimes @xmath127 are presented in fig .
[ flifetimes ] along with calculated term energies expressed relative to the potential minimum of the a state .
the energies are plotted in the order of the values listed in table 1 of ref .
@xcite and correspond to a range of values of @xmath128 from ( 0,15 ) to ( 24,25 ) .
the agreement between @xmath127 and @xmath126 is good within the experimental precision of @xmath129 percent @xcite .
the quasi - bound levels and continua of the @xmath0 state inside or above the high centrifugal potential barriers are found to be important in calculating lifetimes of high @xmath32 levels .
for instance , by including transitions to three quasi - bound levels and the continuum states the calculated lifetime of the @xmath1 @xmath130 level changes from 22.7 ns to 19.3 ns which is close to the measured lifetime of @xmath131 ns @xcite .
the calculated and measured lifetimes agree well up to about @xmath132 @xmath49 in agreement with the findings of ref .
@xcite . from approximately @xmath133 @xmath49 and higher
the experimental lifetimes slightly decrease relative to theory by about 0.8 ns or five percent , as demonstrated in fig .
[ flifetimes ] .
we investigated whether the @xmath134 state might supply an additional spontaneous decay channel , but the theoretical value @xmath135
@xcite for this state appears to place its minimum at around @xmath136 @xmath49 relative to the minimum of the @xmath1 state , apparently ruling this mechanism out .
the reason for the significant downturn of experimental lifetimes for higher term energies is currently not understood ; however , the overall excellent agreement between our calculated lifetimes and the measurements gives us confidence in our molecular data and calculational procedures .
we compare selected examples from the present results with calculations by watson @xcite and by sangfelt _ et al . _
@xcite in table [ t : alife ] .
the calculations of sangfelt _ et al . _ are larger than ours , probably because theoretical transition energies were used .
a simple rescaling using experimental energies , as pointed out by those authors , yields lifetimes in good agreement with the present work .
the dipole moment function calculated by watson @xcite is in good agreement with that adopted in the present study and can not account for the shorter lifetimes obtained in that study .
we present a more extensive tabulation of @xmath1 lifetimes in table [ t : alife2 ] covering the same values tabulated in table vii of ref .
@xcite .
lifetimes for levels of the @xmath2 state have been calculated by uzer , watson , and dalgarno @xcite , uzer and dalgarno @xcite , and sangfelt _ et al . _
@xcite and there appear to be no experimental data . in table [ t : tlifetimes ] we compare our calculated lifetimes of selected @xmath2 levels with available calculations for higher values of @xmath121 .
the present results lie between the calculations of uzer _ et al . _
@xcite and those of sangfelt _ et al . _
the lifetimes calculated by uzer
_ et al . _ are longer than ours because their transition dipole moment function , calculated using a model potential method , is smaller than the _ ab initio _
dipole moment of ratcliff _
et al . _
@xcite adopted in the present study .
the dipole moment function calculated by sangfelt _
et al . _ , on the other hand , is in good agreement with that adopted in the present work . as those authors pointed out , and as our results illustrate , the utilization in their calculation of calculated excitation energies which were larger than experimental energies yielded lifetimes that were too short . in table
[ t : tlifetimes2 ] we present a more extensive tabulation of @xmath2 lifetimes covering the same values tabulated in table viii of ref .
@xcite .
absorption spectra arising from molecular singlet transitions at the far wings of the atomic 2p line at 671 nm consist of a blue wing due to @xmath0-@xmath2 transitions and a red wing due to @xmath0-@xmath1 transitions .
calculations for bound bound ( bb ) and bound free ( bf ) absorption cross sections were carried out separately using eq .
( [ e : lam ] ) with @xmath13 = 8 @xmath49 .
the results for the total ( the sum of bb and bf ) absorption cross sections at temperatures of 1000 k , 1500 k and 2033 k are given in figs .
[ s1000k][s2033k ] . the ratios of the peak cross sections of the @xmath0@xmath2 wing to those of the @xmath0@xmath1 wing are higher at lower temperatures . as temperature increases ,
absorption spectra spread out from the peak spectral regions and there emerges near 900 nm a satellite feature arising from the minimum @xcite in the @xmath0-@xmath1 difference potential and the maximum of the transition dipole moment function @xcite .
we also show the bf contributions to the cross sections on each plot .
the bf component contributes mainly to the extreme blue part of the blue wing and increases in magnitude significantly as the temperature increases .
it is found that transitions to quasibound and continuum levels of the @xmath2 state contribute significantly to the total absorption spectra in the case of @xmath0-@xmath2 transitions , apparently because there is less vibrational oscillator strength density in the discrete part of the spectrum for the b state compared to the a state .
transitions into quasibound states of the @xmath1 or @xmath2 states have been included in the results for the total cross sections in figs .
[ s1000k][s2033k ] .
theoretical quantum - mechanical calculations for absorption cross sections from the @xmath0 state to the @xmath1 state over the spectral range 600950 nm were carried out by lam _
et al . _
@xcite using a constant dipole moment of 6.58 d at 1020 k , for which bf transitions are not significant . our calculations in fig .
[ s1000k ] are about a factor of 10 less than the result shown in fig . 5 of ref .
@xcite , but agree well both in overall shape and in details of finer structures .
we repeated the calculations using the constant dipole moment of ref .
@xcite for both classical and quantum - mechanical cross sections and although these two results agreed with each other , they were also a factor of 10 less than the result shown in fig . 5 of ref .
thus it appears to us that there may be a mislabeling of the vertical axis in fig . 5 of lam _ et al .
_ [ a similar calculation that we performed @xcite for @xmath137 at 800 k is in complete agreement with fig . 4 of ref .
@xcite . ] calculations of absorption spectra at 2033 k over the spectral range 450750 nm presented in fig .
[ s2033k ] are in good agreement with the measured values of erdman _
et al . _
@xcite and with quantum - mechanical calculations performed by mills @xcite .
the experimental study of erdman _
et al . _
@xcite involved an investigation of molecular triplet states @xcite and did not explore the satellite feature at 900 nm .
the calculations over the range 450750 nm by mills @xcite included triplet molecular transitions and are not directly comparable with the present results . nevertheless since the singlet transitions dominate the absorption we find excellent qualitative agreement with the calculations presented by mills .
we thank r. ct for generously sharing assembled data and a. dalgarno for helpful discussions .
we also are grateful to a. gallagher , w. stwalley , and m. fajardo for helpful correspondence .
this work is supported in part by the national science foundation under grant phy97 - 24713 and by a grant to the institute for theoretical atomic and molecular physics at harvard college observatory and the smithsonian astrophysical observatory . | for the lithium dimer we calculate cross sections for absorption of radiation from the vibrational - rotational levels of the ground @xmath0 electronic state to the vibrational levels and continua of the excited @xmath1 and @xmath2 electronic states .
theoretical and experimental data are used to characterize the molecular properties taking advantage of knowledge recently obtained from photoassociation spectroscopy and ultra - cold atom collision studies .
the quantum - mechanical calculations are carried out for temperatures in the range from 1000 to 2000 k and are compared with previous calculations and measurements . |
the study and understanding of heterogeneously catalyzed reactions is a field that contains an enormous wealth of still unclear or even completely unexplained phenomena .
this scenario leads to an exciting and challenging domain for investigation and fundamental research .
furthermore , the occurrence of many complex and fascinating physical and chemical phenomena , such as pattern formation and self - organization @xcite , regular and irregular kinetic oscillations @xcite , propagation and interference of chemical waves and spatio - temporal structures @xcite , the transition into chaotic behaviour @xcite , fluctuation - induced transitions @xcite , irreversible phase transitions ( ipt s ) @xcite , etc , has attracted the attention of many scientists .
in addition to the basic interest , heterogeneous catalysis is a field of central importance for numerous industrial ( e.g. synthesis of ammonia , sulfuric and nitric acids , cracking and reforming processes of hydrocarbons , etc . ) and practical ( e.g. catalytic control of environmental pollution such as the emission of @xmath0 , @xmath1 , @xmath2 , @xmath3 , etc . ) applications .
furthermore , information technology , material science , corrosion , energy conversion , ecology and environmental sciences , etc .
are some fields whose rapid growth is somehow based on the recent progress in the study of heterogeneous reactions occurring on surfaces and interfaces .
it should be noticed that recent developments of experimental techniques such as scanning tunneling microscopy ( stm ) , low energy electron diffraction ( leed ) , high resolution electron energy loss spectroscopy ( hreels ) , ultraviolet photoelectric spectroscopy ( ups ) , photoelectron emission microscopy ( peem ) , etc .
@xcite , just to quote few of them , allows the scientists to gather detailed physical and chemical information about surfaces , adsorbates and reaction products . within this context
, the stm based measurement of the reaction rate parameters at a microscopic level for the catalytic oxidation of @xmath0 @xcite is a clear example of the progress recently achieved .
remarkably , the measured parameters agree very well with those previously obtained by means of macroscopic measurements .
also , all elementary steps of a chemical reaction have been induced on individual molecules in a controlled step - by - step manner with the aid of stm techniques @xcite .
furthermore , very recently , the oxidation of @xmath0 on @xmath4 was studied by means of stm techniques inside a high - pressure flow reactor , i.e. under semirealistic conditions as compared with those prevailing in the actual catalytic process @xcite .
it is interesting to notice that a new reaction mechanism , not observed when the reaction takes place under low pressure , has been identified @xcite . due to this stimulating generation of accurate experimental information ,
the study of catalyzed reaction systems is certainly a challenging scientific field for the development and application of analytical methods , theories and numerical simulations . within this context ,
the aim of this report is to review recent progress in the understanding of ipt s occurring in various lattice gas reaction systems ( lgrs ) .
it should be noticed that lgrs models are crude approximations of the actual ( very complex ) catalytic processes .
however , from the physicist s point of view , the lgrs approach is widely used because it is very useful to gain insight into far - from equilibrium processes .
in fact , due to the lack of a well - established theoretical framework , unlike the case of their equilibrium counterpart , the progress of the statistical mechanics of systems out of equilibrium relies , up to some extent , on the study and understanding of simple models .
so , in the case of lgrs , one renounces to a detailed description of the catalyzed reaction and , instead , the interest is focused on the irreversible critical behaviour of archetype models inspired in the actual reaction systems . keeping these concepts in mind
, the review will be devoted to survey the theoretical development of the field during the last decade and to discuss promising areas for further research as well as open questions . in most cases ,
heterogeneously catalyzed reactions proceed according to well - established elementary steps .
the first one comprises trapping , sticking and adsorption .
gaseous reactant atoms and/or molecules are trapped by the potential well of the surface .
this rather weak interaction is commonly considered as a physisorbed precursor state .
subsequently , species are promoted to the chemisorbed state where a much stronger interaction potential is activated . particularly important from the catalytic point of view
is that molecules frequently undergo dissociation , e.g. @xmath5 , @xmath6 , @xmath7 , etc , which is a process that frees highly reactive atomic species on the surface .
sticking and adsorption processes depend on the surface structure ( both geometric and electronic ) . in some cases , chemisorption of small atoms and molecules may induce the reconstruction of the surface .
this effect , coupled to structure dependent sticking coefficients , may lead to the occurrence of collective phenomena such as oscillations @xcite .
after adsorption , species may diffuse on the surface or , eventually , become absorbed in the bulk . due to collisions between adsorbed species of different kind
the actual reaction step can occur .
of course , this step requires that energetic and spatial constraints be fulfilled .
the result of the reaction step is the formation of a product molecule
. this product can be either an intermediate of the reaction or its final output .
the final step of the whole reaction process is the desorption of the products .
this step is essential not only for the practical purpose of collecting and storing the desired output , but also for the regeneration of the catalytic active sites of the surface .
most reactions have at least one rate limiting step , which frequently makes the reaction prohibitively slow for practical purposes when , for instance , it is intended in an homogeneous ( gas or fluid ) medium .
the role of a good solid - state catalyst is to obtain acceptable output rate of the products .
reactions occurring in this way are commonly known as heterogeneously catalyzed . at this stage and in order to illustrate the above - mentioned elementary steps , it is useful to point our attention to a specific reaction system . for this purpose ,
the catalytic oxidation of carbon monoxide , namely @xmath8 , which is likely the most studied reaction system , has been selected .
it is well known that such reaction proceeds according to the langmuir - hinshelwood mechanism @xcite , i.e with both reactants adsorbed on the catalyst s surface @xmath9 @xmath10 @xmath11 where @xmath12 is an empty site on the surface , while ( @xmath13 ) and ( @xmath14 ) refer to the adsorbed and gas phase , respectively . the reaction takes place with the catalyst , e.g @xmath15 , in contact with a reservoir of @xmath0 and @xmath6 whose partial pressures are @xmath16 and @xmath17 , respectively . equation ( 1 )
describes the irreversible molecular adsorption of @xmath0 on a single site of the catalyst s surface .
it is known that at under suitable temperature and pressure reaction conditions , @xmath0 molecules diffuse on the surface .
furthermore , there is a small probability of @xmath0 desorption that increases as the temperature is raised @xcite
. equation ( 2 ) corresponds to the irreversible adsorption of @xmath6 molecules that involves the dissociation of such species and the resulting @xmath18 atoms occupy two sites of the catalytic surface . under reaction conditions
both the diffusion and the desorption of oxygen are negligible .
due to the high stability of the @xmath6 molecule the whole reaction does not occur in the homogeneous phase due to the lack of @xmath6 dissociation .
so , equation ( 2 ) dramatically shows the role of the catalyst that makes feasible the rate limiting step of the reaction .
finally , equation ( 3 ) describes the formation of the product ( @xmath19 ) that desorbs from the catalyst s surface .
this final step is essential for the regeneration of the catalytic active surface . assuming irreversible adsorption - reaction steps , as in the case of equations ( 1 - 3 )
, it may be expected that on the limit of large @xmath16 and small @xmath17 ( small @xmath16 and large @xmath17 ) values , the surface of the catalyst would become saturated by @xmath0 ( @xmath6 ) species and the reaction would stop .
in fact , the surface of the catalyst fully covered by a single type of species , where further adsorption of the other species is no longer possible , corresponds to an inactive state of the system .
this state is known as ` poisoned ' , in the sense that adsorbed species on the surface of the catalyst are the poison that causes the reaction to stop .
physicists used to call such state ( or configuration ) ` absorbing ' because a system can be trapped by it forever , with no possibility of escape @xcite .
these concepts are clearly illustrated in figure 1 , which shows plots of the rate of @xmath19 production ( @xmath20 ) and the surface coverage with @xmath0 and @xmath6 ( @xmath21 and @xmath22 , respectively ) , versus the partial pressure of @xmath0 ( @xmath16 ) , as obtained using the ziff - gulari - barshad ( zgb ) lattice gas reaction model @xcite .
details on the zgb model will be discussed extensively below , see also @xcite . for @xmath23
the surface becomes irreversibly poisoned by @xmath18 species with @xmath24 , @xmath25 and @xmath26 .
in contrast , for @xmath27 the catalyst is irreversibly poisoned by @xmath0 molecules with @xmath28 , @xmath29 and @xmath26 .
these poisoned states are absorbing and the system can not escape from them .
however , as shown in figure 1 , between these absorbing states there is a reaction window , namely for @xmath30 , such that a steady state with sustained production of @xmath19 is observed .
it is worth mentioning that starting from the reactive regime and approaching the oxygen absorbing state , all quantities of interest change smoothly until they adopt the values corresponding to the absorbing state .
this behaviour typically corresponds to a second - order irreversible phase transition ( ipt ) .
the transition is irreversible because when the control parameter ( @xmath16 in this example ) is tuned into the absorbing state the system becomes trapped by it forever .
this behaviour is in contrast to that observed for second - order reversible phase transitions , such as the order - disorder transition of the ising ferromagnet in the absence of an external magnetic field , where it is possible to change reversibly from one phase to the other , simply tuning the control parameter @xcite . for second - order ipt s , as in the case of their reversible counterparts
, it is possible to define an order parameter , which for the former is given by the concentration of minority species ( @xmath21 , in the case of the second - order ipt of the catalytic oxidation of @xmath0 ) .
furthermore , it is known that @xmath21 vanishes according to a power law upon approaching the critical point @xcite , so that @xmath31 where @xmath32 is the order parameter critical exponent and @xmath33 is the critical point .
remarkably , the behaviour of the system is quite different upon approaching the @xmath0 absorbing state from the reactive regime ( see figure 1 ) . in this case , all quantities of interest exhibit a marked discontinuity close to @xmath34 .
this is a typical first - order ipt and @xmath35 is the coexistence point .
experimental results for the catalytic oxidation of carbon monoxide on @xmath36 and @xmath37 @xcite are in qualitative agreement with simulation results of the zgb model , it follows from the comparison of figures 1 and 2 .
a remarkable agreement is the ( almost ) linear increase in the reaction rate observed when the @xmath0 pressure is raised and the abrupt drop of the reactivity when a certain ` critical ' pressure is reached . in spite of
the similarities observed , two essential differences are worth discussing : i ) the oxygen - poisoned phase exhibited by the zgb model within the @xmath0 low - pressure regime is not observed experimentally .
therefore , one lacks experimental evidence of a second - order ipt .
ii ) the @xmath38rich phase exhibiting low reactivity found experimentally resembles the @xmath38poisoned state predicted by the zgb model .
however , in the experiments the nonvanishing @xmath38desorption probability prevents the system from entering into a truly absorbing state and the abrupt , ` first - order like ' transition , shown in figure 2 is actually reversible .
of course , these and other disagreements are not surprising since the lattice gas reaction model , with a single parameter , is a simplified approach to the actual catalytic reaction that is far more complex .
1.5 true cm at this stage it is worth warning the reader that due to the non - hamiltonian nature of the zgb model , as well as all the lattice gas reaction systems that will be treated hereafter , one lacks a thermodynamic quantity , such as the free energy , in order to draw a sharper definition of the order of the transitions .
even more , in contrast to their reversible counterpart , the field of ipt s lacks a well - established theoretical framework . also , it is worth mentioning that systems with absorbing states are clearly out of equilibrium .
in fact , the transition rate out of the absorbing state is zero , such as those configurations can not fulfill standard detailed balance requirements .
therefore , the study of the critical behaviour of these systems must involve ipt s between reactive and absorbing phases . due to these circumstances ,
the study of ipt s represents a theoretical challenge within the more general goal of modern physics given by the development of the statistical mechanics of nonequilibrium systems .
it should be recognized that the study of absorbing states and ipt s is almost ignored in the courses of statistical mechanics , in spite of the fact that they abound in physics , chemistry , biology and other disciplines including sociology .
some typical examples include the spreading of epidemics through a population , the propagation of rumors in a society @xcite , the spreading of fire through a forest @xcite , coexistence with extinction transitions in prey - predator systems @xcite and , of course , catalytic and autocatalytic chemical reactions . from a more general point of view
, absorbing states are expected to occur in situations where some quantity of interest can proliferate or die out ( e.g the fire in the forest ) , without any possibility of spontaneous generation ( e.g due to the rays of an electrical storm ) .
the underlying physics involves the competition between proliferation and death of a certain quantity .
proliferation is associated with the active ( reactive ) phase of the system , while inactivity ( poisoning ) characterizes the absorbing phase .
there are currently three basic approaches for the theoretical modeling of surface reactions : i ) _ ab - initio _ calculations , ii ) analytic approaches and iii ) stochastic models .
the _ ab - initio _ method is usually implemented via density functional theory approaches and due to the huge computational requirements , the calculations are restricted to few atomic layers of the catalysts and a very reduced catalyst s surface ( of the order of @xmath39 ) .
this approach frequently allows the study of a single adsorbed species or a reactive pair only .
consequently , the study of macroscopic surface systems incorporating statistical effects , as requested for the case of critical phenomena , is impossible at present .
so , this approach will not be further discussed here . on the other hand ,
stochastic models can account for fluctuations in large systems .
so , they are used to deal with a large number of collective phenomena occurring in reaction systems that are not only restricted to ipt s , but also involve spatio - temporal structures @xcite , chemical waves @xcite , kinetic oscillations @xcite , the transition to chaos @xcite , etc .. broad experience gained in the treatment of equilibrium systems has shown that monte carlo simulations @xcite and renormalization group ( rg ) analysis of classical field - theoretical models @xcite are among the most useful tools for the treatment of phase transitions and critical phenomena .
a much more reduced experience , obtained during the last decade , indicates that , after some adaptation , similar techniques can also been employed to deal with ipt s .
monte carlo ( mc ) simulations of heterogeneously catalyzed reactions can be considered the computational implementation of microscopic reaction mechanisms .
in fact , such mechanisms are the ` rules ' of the computer algorithm .
of course , the operation of the rules may lead to the development of correlations , while stochastic fluctuations are inherent to the method . for the practical implementation of the mc method , the catalyst s surface is replaced by a lattice .
therefore , lattice gas reaction models are actually considered .
for this reason , the method often faces the limitations imposed by the size of the lattices used . in some particular cases ,
e.g. when studying second - order phase transitions , this shortcoming can be overcome appealing to the well - established finite - size - scaling theory @xcite . also , very often one can develop extrapolation methods that give reliable results for the thermodynamic limit , i.e. infinite lattices .
another limitation arises when the diffusion rate of the adsorbed species is very large . in this case , most of the computational time has to be devoted to the diffusion process while the quantity of interest , namely the number of reaction events , becomes negligible .
this drawback may be overcome implementing a mixed treatment : mean - field description of the diffusion and mc simulation of the reaction @xcite .
this approach may become an interesting and powerful tool in the near future .
mc simulations of dynamic and kinetic processes are often hindered by the fact that the monte carlo time , usually measured in monte carlo time steps , is somewhat proportional to the actual time .
so , direct comparison with experiments becomes difficult .
however , very recently a more sophisticated implementation of the mc method has been envisioned : namely the dynamic monte carlo ( dmc ) approach that incorporates the actual time dependence of the processes allowing direct comparison with experiments @xcite .
further developments and applications of the dmc method are a promising field of research . within this context ,
the following subsections describe different mc approaches suitable for the study of ipt s in reaction systems , namely the standard ensemble , the constant coverage ensemble and the epidemic method .
furthermore , the standard finite - size scaling theory adapted to the case of second - order ipt s and they application to mc data are also discussed .
in this ensemble the catalyst is assumed to be in contact with an infinitely large reservoir containing the reactants in the gas phase .
adsorption events are treated stochastically neglecting energetic interactions .
the reaction between reactants takes place on the surface of the catalyst , i.e. the so - called langmuir - hinshelwood mechanism .
after reaction , the product is removed from the surface and its partial pressure in the gas phase is neglected , so that readsorption of the product is not considered . in order to further illustrate the practical implementation of the standard ensemble
let us describe the simulation method of the lattice gas reaction version of the catalytic oxidation of @xmath0 ( equations ( 1 - 3 ) ) according to the zgb model @xcite on the two - dimensional square lattice .
the monte carlo algorithm is as follows : \i ) @xmath0 or @xmath6 molecules are selected at random with relative probabilities @xmath16 and @xmath40 , respectively .
these probabilities are the relative impingement rates of both species , which are proportional to their partial pressures in the gas phase in contact with the catalyst . due to the normalization , @xmath41
, the model has a single parameter , i.e. @xmath16 . if the selected species is @xmath0 , one surface site is selected at random , and if that site is vacant , @xmath0 is adsorbed on it according to equation ( 1 ) . otherwise , if that site is occupied , the trial ends and a new molecule is selected .
if the selected species is @xmath6 , a pair of nearest - neighbor sites is selected at random and the molecule is adsorbed on them only if they are both vacant , as requested by equation ( 2 ) .
\ii ) after each adsorption event , the nearest - neighbors of the added molecule are examined in order to account for the reaction given by equation ( 3 ) .
if more than one [ @xmath42 pair is identified , a single one is selected at random and removed from the surface .
the phase diagram of the zgb model , as obtained using the standard ensemble , is shown in figure 1 and will be further discussed in .
monte carlo simulations using the constant coverage ( cc ) ensemble , as early proposed by ziff and brosilow @xcite , are likely the most powerful method available for the study of first - order ipt s . in order to implement the cc method ,
first a stationary configuration of the system has to be achieved using the standard ensemble algorithm , as described in for this purpose , one selects a value of the parameter close to the coexistence point but within the reactive regime . after achieving the stationary state ,
the simulation is actually switched to the cc method .
then , one has to maintain the coverage with the majority species of the corresponding absorptive state as constant as possible around a prefixed value .
this goal is achieved regulating the number of adsorption events of competing species .
the ratio between the number of attempts made trying to adsorb a given species and the total number of attempts is the measure of the corresponding effective partial pressure . in this way
, in the cc ensemble the coverage assumes the role of the control parameter . in the case of the zgb model , the stationary state is usually achieved for @xmath43 .
so , in order to maintain the @xmath0 coverage close to the desired value @xmath44 , oxygen ( @xmath0 ) adsorption attempts take place whenever @xmath45 ( @xmath46 ) .
let @xmath47 and @xmath48 be the number of carbon monoxide and oxygen attempts , respectively .
then , the value of the ` pressure ' of @xmath0 in the @xmath49 ensemble ( @xmath50 ) is determined just as the ratio @xmath51 .
subsequently , the coverage is increased by small amount , say @xmath52 . a transient period @xmath53
is then disregarded for the proper relaxation of the system to the new @xmath44 value , and finally averages of @xmath50 are taken over a certain measurement time @xmath54 . in the original cc algorithm of brosilow and ziff @xcite the coverage of @xmath0
was increased stepwise up to a certain maximum value ( @xmath55 ) and the set of points ( @xmath56 ) were recorded .
however , later on it has been suggested @xcite that it would be convenient to continue the simulations after reaching @xmath55 by * decreasing * @xmath44 stepwise until @xmath50 reaches a value close to the starting point , namely @xmath57 .
in fact , this procedure would allow to investigate one possible hysteretic effects at first - order ipt s , which are expected to be relevant as follows from the experience gained studying their counterpart in equilibrium ( reversible ) conditions .
since true phase transitions only occur on the thermodynamic limit and computer simulations are always restricted to finite samples , numerical data are influenced by rounding and shifting effects around pseudo critical points @xcite . within this context
the finite - size scaling theory @xcite has become a powerful tool for the analysis of numerical results allowing the determination of critical points and the evaluation of critical exponents .
all this experience gained in the study of reversible critical phenomena under equilibrium conditions can be employed in the field of second - order ipt s . in order to perform a finite - size scaling analysis close to a second - order itp @xcite in a reaction system
it is convenient to take the concentration of the minority species on the surface , ( @xmath58 ) , as an order parameter .
by analogy to reversible second - order transitions on assumes that @xmath59 where @xmath32 is the order parameter critical exponent and @xmath60 is the critical value of the control parameter @xmath61 .
approaching @xmath60 from the reactive regime , the characteristic length scale of the system given by the correlation length @xmath62 diverges according to @xmath63 where @xmath64 is the correlation length exponent in the space direction . for finite samples and close to the critical region , the concentration of minority species will depend on two competing lengths , namely @xmath65 , and the scaling hypothesis assumes @xmath66 , \ ] ] where
equation ( [ corlen ] ) has been used and @xmath67 is a suitable scaling function . just at @xmath60 , one has @xmath68 and @xmath69 such that equation ( [ orpa ] ) is recovered when @xmath70 in the critical region . as anticipated by the adopted notation , second - order ipt s exhibit spatio - temporal anisotropy , so that the correlation length in the time direction is given by @xmath71 where @xmath72 is the corresponding correlation length exponent .
examples of the application of finite - size scaling to various reaction systems can be found in the literature @xcite .
however , a more accurate method for the evaluation of critical exponents is to combine finite - size scaling of stationary quantities and dynamic scaling , as will be discussed just below . in order to obtain accurate values of the critical point and the critical exponents using the standard ensemble and applying finite - size scaling
, it is necessary to perform mc simulations very close to the critical point .
however , at criticality and close to it , due to the large fluctuations always present in second - order phase transitions , any finite system will ultimately become irreversibly trapped by the absorbing state .
so , the measurements are actually performed within metastable states facing two competing constraints : on the one hand the measurement time has to be long enough in order to allow the system to develop the corresponding correlations and , on the other hand , such time must be short enough to prevent poisoning of the sample .
this shortcoming can be somewhat healed by taking averages over many samples and disregarding those that have been trapped by the absorbing state .
however , it is difficult to avoid those samples that are just evolving to such absorbing state , unless that fluctuations are suppressed by comparing two different samples evolving through the phase space following very close trajectories @xcite . in view of these shortcomings
, experience indicates that the best approach to second - order ipt s is to complement finite - size scaling of stationary quantities , as obtained with the standard ensemble , with epidemic simulations .
the application of the epidemic method ( em ) to the study of ipt s has become a useful tool for the evaluation of critical points , dynamic critical exponents and eventually for the identification of universality classes @xcite .
the idea behind the em is to initialize the simulation using a configuration very close to the absorbing state .
such a configuration can be achieved generating the absorbing state using the standard ensemble and , subsequently , removing some species from the center of the sample where a small patch of empty sites is left . in the case of the zgb model
this can be done by filling the whole lattice with @xmath0 , except for a small patch .
patches consisting of 3 - 6 neighboring empty sites are frequently employed , but it is known that the asymptotic results are independent of the size of the initial patch .
such patch is the kernel of the subsequent epidemic .
after the generation of the starting configuration , the time evolution of the system is followed using the standard ensemble as already described in . during this dynamic process
the following quantities are recorded : ( i ) the average number of empty sites ( @xmath73 ) , ( ii ) the survival probability @xmath74 , which is the probability that the epidemic is still alive at time @xmath75 , and ( iii ) the average mean square distance , @xmath76 , over which the empty sites have spread . of course ,
each single epidemic stops if the sample is trapped in the poisoned state with @xmath77 and , since these events may happen after very short times ( depending on the patch size ) , results have to be averaged over many different epidemics
. it should be noticed that @xmath73 ( @xmath78 ) is averaged over all ( surviving ) epidemics .
if the epidemic is performed just at critically , a power - law behaviour ( scaling invariance ) can be assumed and the following anstze are expected to hold , @xmath79 @xmath80 and @xmath81 where @xmath82 , @xmath83 and @xmath84 are dynamic critical exponents .
thus , at the critical point log - log plots of @xmath73 , @xmath74 and @xmath78 will asymptotically show a straight line behaviour , while off - critical points will exhibit curvature .
this behaviour allows the determination of the critical point and from the slopes of the plots the critical exponents can also be evaluated quite accurately @xcite .
using scaling arguments , it has been shown that the following relationship @xcite @xmath85 holds , allowing the evaluation of one exponent as a function of the other two .
the validity of equations ( [ nu ] ) , ( [ so ] ) and ( [ r2 ] ) for second - order ipt s is very well established .
furthermore , the observation of a power - law behaviour for second - order ipt s is in agreement with the ideas developed in the study of equilibrium ( reversible ) phase transitions : scale invariance reflects the existence of a diverging correlation length at criticality .
it should be mentioned that the em can also be applied to first - order ipt s close to coexistence . however , since it is well known that in the case of first - order reversible transitions correlations decay exponentially , preventing the occurrence of scale invariance , equations ( [ nu ] ) , ( [ so ] ) and ( [ r2 ] ) have to be modified .
recently , the following anzats has been proposed @xcite @xmath86 \label{anz}\ ] ] where @xmath87 is an effective exponent and @xmath88 sets a characteristic time scale .
so , equation ( [ anz ] ) combines a power - law behaviour for @xmath89 with an exponential ( asymptotic ) decay .
it should also be mentioned that the whole issue of the occurrence of power - law behaviour at equilibrium first - order transitions is a bit more complex than the simplified arguments used above .
for example , scaling behaviour inside the coexistence zone has been observed for the liquid - gas phase transitions using finite systems .
however , this scaling disappears on the thermodynamic limit @xcite .
also , when a first - order line ends in a second - order critical point , the system frequently displays several decades of critical behaviour ( before the exponential roll - off ) even when measurements are performed quite a bit beyond the critical point @xcite . in the field of reversible critical phenomena ,
the most effective analytical tool for the identification of universality classes is the renormalization group analysis of classical field - theoretical models using coarse - grained ginzburg - landau - wilson free - energy functionals @xcite .
while reaction systems exhibiting ipt s do not have energy functionals , they can often be treated on the coarse - grained level by means of phenomenological langevin equations .
stochastic partial differential equations of this kind are the natural formalism to analyze critical properties of irreversible systems with absorbing states , as will be discussed below . on the other hand , mean - field modeling using ordinary differential equations ( ode )
is a widely used method @xcite for the study of first - order ipt s ( see also below ) .
further extensions of the ode framework , to include diffusional terms are very useful and , have allowed the description of spatio - temporal patterns in diffusion - reaction systems @xcite . however , these methods are essentially limited for the study of second - order ipt s because they always consider average environments of reactants and adsorption sites , ignoring stochastic fluctuations and correlations that naturally emerge in actual systems .
the langevin equation for a single - component field @xmath90 can be written as @xcite @xmath91 where @xmath92 and @xmath93 are functionals of @xmath94 , and @xmath82 is a gaussian random variable with zero mean , such that the only nonvanishing correlations are @xmath95 .
equation ( [ lan ] ) is purely first - order in time since it is aimed to describe critical properties where small - wavenumber phenomena dominate the physical behaviour .
the goal of the coarse - grained lagevin representation is to capture the essential physics of microscopic models on large length and time scales .
therefore , @xmath92 and @xmath93 are taken to be analytical functions of @xmath90 and its spatial derivatives . in this way ,
coarse graining smooths out any nonanalyticity of the underlying microscopic dynamic model . at this stage , it is clear that one has an infinite number of possible analytical terms in @xmath94 and its space derivatives that can be placed on the right hand side term of equation ( [ lan ] ) .
following the ideas of landau and lifshitz , it is expected that symmetry considerations would suffice to determine the relevant terms of @xmath92 and @xmath93 .
in fact , these functional must include all analytic terms consistent with the symmetry of the microscopic model and no term with lower symmetry can be included .
the coefficients of these terms are simply taken as unknown parameters or constants to be determined phenomenologically .
this assumption is consistent with the fact that the typical ipt s , wich are intended to be described with equation ( [ lan ] ) , can be characterized by a set of critical exponents defining their universality class , which do not depend on the values of the parameters .
depending on the behaviour of the noise amplitude functional @xmath93 , two types of physical systems can already be treated : \a ) systems without absorbing states . in this case @xmath96
approaches a constant value . using renormalization group arguments
it can be shown that higher powers of @xmath94 and its derivatives in @xmath93 are irrelevant for the description of the critical behaviour and therefore , they can be neglected . in this way
, the langevin equation has a simple additive gaussian noise of constant amplitude . assuming that @xmath97 can be written as @xmath98 where @xmath99 and @xmath100 are constants , the best known example results in the ginzburg - landau - wilson functional derivative of free energy functional from the ising model , e.g. @xmath101,\ ] ] such that the langevin equation becomes the celebrated time dependent ginzburg - landau equation
this equation is very useful for the treatment of equilibrium critical phenomena and has been quoted here for the sake of completeness only .
in fact , in this article , our interest is focused on far - from equilibrium systems , for which @xmath97 can not be expressed as a functional derivative . within this context
the best known example is the kardar - parisi - zhang ( kpz ) equation introduced for the description of the dynamic behaviour of interfaces without overhangs . for systems with interfaces , @xmath102 is taken as the height at time @xmath75 of a site of the interface at position @xmath103 .
it has been shown that the functional @xmath104 captures the large scale physics of moving interfaces leading to the kpz equation @xcite @xmath105 recent studies have demonstrated that the kpz equation holds for the treatment of a wide variety of physical situations @xcite including the description of the interfacial behaviour of reactants close to coexistence in systems exhibiting first - order ipt s @xcite , for further details see also .
\b ) systems with absorbing states . in the langevin equations capable of describing these systems ,
the noise amplitude functional @xmath96 must vanish with @xmath94 so that the state @xmath106 is an absorbing state for any functional @xmath92 that has no constant piece and therefore it also vanishes with @xmath94 .
it can be demonstrated that @xmath107 , for small enough values of @xmath94 . also , @xmath108 can assume only positive integer and half - odd - integer values .
the cases @xmath109 and @xmath110 are those of interest in practice , the former includes the directed percolation ( dp ) universality class ( which is of primary interest in the context of this article because it is suitable for the description of second - order ipt s @xcite ) , while the latter includes the problem of multiplicative noise . considering the case of dp @xcite , one has to keep in mind that the state @xmath111 must be absorbing , so that both functionals @xmath97 and @xmath96 must vanish as @xmath112 .
this constraint implies that terms independent of @xmath94 must can not be considered .
now , imposing the condition of symmetry inversion on the underlying lattice ( @xmath113 ) , so that terms containing @xmath114 are forbidden , the langevin equation with the lowest allowed terms in @xmath94 and its derivatives becomes @xmath115 where @xmath116 , @xmath99 and @xmath100 are constants .
using renormalization group methods it can be shown that the critical dimension for equation ( [ lan1 ] ) is @xmath117 and that the critical exponents in terms of the first order @xmath118expansion ( here @xmath119 ) are @xcite : dynamic exponent @xmath120 , order parameter critical exponent @xmath121 and correlation length exponent @xmath122 . also , the scaling relation @xmath123 holds , where @xmath124 is the critical exponent of the two - point correlation function .
therefore , only two of the exponents are truly independent .
notice that equation ( [ scarel ] ) is a nonequilibrium counterpart of the well - known scaling relation @xmath125 valid for a system without absorbing states .
apart from the scaling relations given by equation ( [ anzatgdt ] ) and ( [ scarel ] ) , which hold for dynamic and static exponents , the following relationships between exponents have been found to hold @xcite : @xmath126 , @xmath127 and @xmath128 .
it should also be noticed that ipt s subject to extra symmetries , and thus out of the dp universality class , have been identified in recent years . among
then one has systems with symmetric absorbing states @xcite , models of epidemics with perfect immunization @xcite , and systems with an infinite number of absorbing states @xcite .
additional discussions on these topics are beyond the aim of this article since these situations do not appear in the reaction models discussed here .
for further details the reader is addressed to the recent review of hinrichsen @xcite .
the mean field ( mf ) approach to irreversible reactions neglects spatial fluctuations and the effect of noise .
so , the actual concentrations of the reactants are replaced by averaged values over the sites of the lattice .
differential equations for the time dependence of such averages are derived for specific reaction systems with different degrees of sophistication , for instance involving one or more site correlations .
mf equations are useful in order to get insight into the behaviour of reaction systems close to first - order ipt s where spatial fluctuations are expected to be irrelevant @xcite . in spite of the restricted application of mf equations to the study of second - order ipt s , it is instructive to discuss the mf predictions of the langevin equation ( [ lan1 ] ) .
in fact , replacing the field @xmath102 by the spatial constant @xmath129 and neglecting the noise and the laplacian terms , equation ( [ lan1 ] ) becomes @xcite @xmath130 equation ( [ lanmf ] ) has two long - time stationary solutions given by the absorbing state ( @xmath111 ) and the active regime ( @xmath131 ) , wich are stable for @xmath132 and @xmath133 . here
@xmath99 plays the role of the control parameter with a critical point at @xmath134 .
the order parameter critical exponent that governs the decay of the average density @xmath135 for @xmath136 from the active phase is @xmath137 .
the first part of this section will be devoted to describe lattice gas reaction models inspired in actual catalytic reactions , such as the catalytic oxidation of carbon monoxide ( ) and the reaction between nitric oxide and carbon monoxide ( ) .
these models exhibit many features characteristic of first - order ipt s that have also been found in numerous experiments , such as hysteretic effects @xcite and abrupt changes of relevant properties when a control parameter is tuned around a coexistence point @xcite ( see also figure 2 ) , etc . on view of these facts
, special attention will be drawn to the discussion of first - order itp s .
on the other hand , will be mostly devoted to the discussion of generic models , which are not intended to describe specific reaction systems .
most of these models exhibit second - order itp s . as already discussed in the introductory sections ,
the catalytic oxidation of carbon monoxide is one of the most studied reaction due to its practical importance and theoretical interest . the simplest approach for this reaction is the zgb lattice gas model @xcite as described in and ( for additional details see @xcite ) .
the phase diagram of the zgb model ( figure 1 ) exhibits a second - order ipt close to @xmath138 that belongs to the dp universality class @xcite .
the dynamic critical exponents , as evaluated using epidemic simulations ( ) , equations ( [ nu ] ) , ( [ so ] ) and ( [ r2 ] ) ) , are @xmath139 , @xmath140 and @xmath141 ( in two dimensions ) @xcite , which , in fact , are in excellent agreement with the acepted exponents of the universality class of the directed percolation , namely @xmath142 , @xmath143 and @xmath144 @xcite .
the order parameter critical exponent , as evaluated using the damage spreading technique @xcite , is @xmath145 , also in excellent agreement with the dp value , namely @xmath146 @xcite .
more interesting , close to @xmath147 , the zgb model exhibits a first - order ipt ( see figure 1 ) in qualitative agreement with experiments performed using single crystal surfaces as catalysts ( see figure 2 ) . as mentioned above , the nonvanishing @xmath0 desorption rate experimentally observed @xcite prevents the actual catalyst system from entering in to a truly absorbing state and the abrupt transition shown in figure 2 is actually reversible .
further experimental evidence on the existence of this first - order - like behaviour arises from dynamic measurements exhibiting clear hysteretic effects , as shown in figure 3 for the case of the @xmath148 reaction on a @xmath37 single crystal surface @xcite . , keeping the oxygen pressure constant at @xmath149 , while the @xmath0 partial pressure is varied cyclically ( horizontal axis ) .
( a ) hysteresis in the reactant coverage as measured by the photoelectron emission microscopy ( peem ) @xcite and ( b ) in the @xmath19 reaction rate .
more details in the text . adapted from reference @xcite.,width=226 ] , keeping the oxygen pressure constant at @xmath149 , while the @xmath0 partial pressure is varied cyclically ( horizontal axis ) .
( a ) hysteresis in the reactant coverage as measured by the photoelectron emission microscopy ( peem ) @xcite and ( b ) in the @xmath19 reaction rate .
more details in the text . adapted from reference @xcite.,width=226 ]
figure 3(a ) shows hysteresis in the reactant coverage upon cyclic variation of the @xmath0 partial pressure ( horizontal axis ) .
the vertical axis shows the photocurrent measured in a photoelectron emission microscopy ( peem ) @xcite experiment .
notice that a low ( negative ) photocurrent indicates an oxygen - rich phase ( left - hand side of figure 3(a ) ) , while a large photocurrent corresponds to a @xmath38rich phase ( right - hand side of figure 3(a ) ) .
also , figure 3(b ) shows the hysteresis in the rate of @xmath19 production ( measured using a mass spectrometer ) as a function of the @xmath0 partial pressure .
when the system is in the low @xmath0 pressure regime , the surface is mostly covered by oxygen that corresponds to the ` oxygen side ' or monostable region a. upon increasing @xmath16 the reaction rate also rises until , close to @xmath150 , the surface becomes covered with adsorbed @xmath0 .
this ` @xmath0 side ' or monostable state b , corresponds to a low reactivity regime . decreasing @xmath16 ,
the system remains in this monostable region b until , close to @xmath151 , it suddenly undergoes a steep transition and rapidly returns to the initial state of high reactivity . summing up ,
during the hysteresis loop the system may be in two monostable regions a ( @xmath152 ) and b ( @xmath153 ) , separated from each other by a bistable region ( @xmath154 ) @xcite . in view of this stimulating experimental evidence
let us review some numerical studies on hysteretic effects close to coexistence .
figure 4 shows a plot of @xmath155 versus @xmath16 obtained by means of the @xmath49 ensemble applied to the zgb model and using a relatively small sample ( @xmath156 ) .
starting from the stationary value of @xmath21 at @xmath157 , one observes that stepwise increments of @xmath21 cause @xmath16 to increase steadily up to the @xmath158-dependent upper spinodal point @xmath159 . turning around the spinodal point , further increments of @xmath21
cause @xmath16 to decrease steadily up to @xmath160 for @xmath161 . at this point
the growing @xmath0 branch finishes and the subsequent decrease in @xmath21 causes the system to return to the starting point where the decreasing @xmath0 branch of the loop ends .
notice that both the growing and decreasing branches describe the same trajectory ( within error bars ) on the ( @xmath162 ) plane .
so , it has been concluded that hysteretic effects are not observed using this small lattice @xcite . increasing the size of the lattice ( @xmath163 in figure 4(b ) ) , the behaviour of the system changes dramatically . on the one hand ,
the spinodal point becomes appreciably shifted ( see inset of figure 4(b ) ) , and on the other hand , hysteretic effects become evident since @xmath0-growing and @xmath0-decreasing branches can be clearly distinguished .
notice that within a remarkably wide range of @xmath0 values both branches are vertical , and consequently parallel each to other . after increasing the lattice size ( @xmath164 in figure 4(c ) ) only minor changes in @xmath165 , @xmath166 , @xmath167 occur , but a well defined spinodal point and a hysteresis loop can still be observed ( see inset of figure 4(c ) ) .
figure 5 shows a plot of the @xmath158-dependent spinodal points ( @xmath168 ) versus the inverse lattice size ( @xmath169 ) . performing an extrapolation to the infinite size limit yields @xmath170 @xcite .
this figure should be compared with the value reported by brosilow @xcite , @xmath171 , which corresponds to a determination of @xmath165 in a finite lattice of size @xmath172 . also , evans @xcite have reported @xmath173 for a finite lattice of size @xmath174 .
very recently , an independent estimation given by @xmath175 , which is in excellent agreement with the figures measured using the cc method , has been obtained by means of short - time dynamic measurements @xcite .
plots of both @xmath166 and @xmath167 versus @xmath176 , also shown in figure 5 , indicate that finite - size effects are negligible for large lattices ( @xmath177 ) .
so , the extrapolated estimates are @xcite @xmath178 and @xmath179 , respectively .
another interesting feature of @xmath49 simulations showing hysteresis is that one can exchange the role of the axis in the following sense : finely tuning the coverage @xmath21 one can induce the system to undergo first - order transitions in parameter space ( @xmath16 in this case ) .
hysteretic effects can be further understood after examination of various snapshot configurations as shown in figure 6 .
it should be noticed that all of the configurations belong to the coexistence region and , consequently , these states are not allowed when simulating the zgb model using the standard algorithm since right at @xmath180 , @xmath21 displays a discontinuous jump from @xmath181 to @xmath182 ( see figure 1 ) . figure 6(a ) shows a typical configuration corresponding to the spinodal point @xmath183 with @xmath184 . here
, one observes that some small but compact @xmath0 clusters have already been nucleated .
this configuration is different from those obtained within the reactive regime using the standard algorithm ( not shown here for the sake of space ) that show mostly @xmath0 monomers with @xmath181 .
the snapshot configurations shown in figures 6(b ) and ( c ) correspond to the growing branch and have been obtained for @xmath185 and @xmath186 , respectively .
it should be noticed that above @xmath165 a single massive @xmath0 cluster has spread within the reactive phase . in figure 6(b ) , this massive @xmath0 cluster does not percolate , but increasing @xmath21 percolation along a single direction of the lattice is observed ( figure 6(c ) ) .
percolation of the massive cluster along only one direction is observed up to a relatively high @xmath0 coverage ( @xmath187 in figure 6(d ) ) .
these values of the @xmath21 are remarkably greater than the percolation threshold of random percolation model , given by @xmath188 @xcite .
however , the random percolation cluster is a fractal object with fractal dimension @xmath189 @xcite , while the @xmath0 cluster is a compact object . dangling ends emerging from the surface of the @xmath0 cluster eventually get in contact causing percolation of such a cluster in both directions of the lattice ( figure 6(e ) ) .
it should be noticed that the snapshot configuration of figure 6(d ) corresponds to the growing branch while that of figure 6(e ) , which has been taken after few mcs , corresponds to an effective @xmath0 pressure characteristic of the decreasing branch .
therefore , the jump from one branch to the other seems to be accompanied by a change in the configuration of the @xmath0 cluster .
it will be interesting to quantitatively study the properties of the interface between the @xmath0 cluster and the reactive phase in order to determine their possible self - affine nature , as well as the interplay between curvature and hysteresis . from the qualitative point of view
, the examination of snapshots suggests that both the interface roughness and length of the massive @xmath0 cluster remain almost unchanged for the growing branch .
when @xmath21 is further increased , the jump to the decreasing branch is eventually characterized by the onset of percolation along both directions of the lattice and , consequently , the macroscopic length and the curvature of the interface may change .
so , the subtle interplay of interfacial properties , such as length , roughness , curvature , etc . has to be studied in detail in order to fully understand the hysteresis loop observed using the @xmath49 ensemble .
it is expected that the longer the interface length of the growing branch the easier the @xmath190 reaction , so one needs a higher effective @xmath0 pressure to keep the coverage constant .
in contrast , along the decreasing branch , the shorter length of the interface inhibits reactions so that a greater oxygen pressure ( smaller @xmath0 pressure ) is needed to achieve the desired @xmath0 coverage .
so , these arguments may explain the existence of two branches .
0.50 true cm , @xmath0-occupied sites are black while other sites are left white ; a ) snapshot obtained at the spinodal point with @xmath191 , b ) , c ) and d ) are snapshots obtained along the growing branch with @xmath185 , @xmath192 , and @xmath193 , respectively ; e ) snapshot obtained few monte carlo steps after figure d ) but now the system has jumped to the decreasing branch with @xmath194.,width=188 ] , @xmath0-occupied sites are black while other sites are left white ; a ) snapshot obtained at the spinodal point with @xmath191 , b ) , c ) and d ) are snapshots obtained along the growing branch with @xmath185 , @xmath192 , and @xmath193 , respectively ; e ) snapshot obtained few monte carlo steps after figure d ) but now the system has jumped to the decreasing branch with @xmath194.,width=188 ] , @xmath0-occupied sites are black while other sites are left white ; a ) snapshot obtained at the spinodal point with @xmath191 , b ) , c ) and d ) are snapshots obtained along the growing branch with @xmath185 , @xmath192 , and @xmath193 , respectively ; e ) snapshot obtained few monte carlo steps after figure d ) but now the system has jumped to the decreasing branch with @xmath194.,width=188 ] 1.0 true cm , @xmath0-occupied sites are black while other sites are left white ; a ) snapshot obtained at the spinodal point with @xmath191 , b ) , c ) and d ) are snapshots obtained along the growing branch with @xmath185 , @xmath192 , and @xmath193 , respectively ; e ) snapshot obtained few monte carlo steps after figure d ) but now the system has jumped to the decreasing branch with @xmath194.,width=188 ] , @xmath0-occupied sites are black while other sites are left white ; a ) snapshot obtained at the spinodal point with @xmath191 , b ) , c ) and d ) are snapshots obtained along the growing branch with @xmath185 , @xmath192 , and @xmath193 , respectively ; e ) snapshot obtained few monte carlo steps after figure d ) but now the system has jumped to the decreasing branch with @xmath194.,width=188 ] the snapshot configurations of figure 5 unambiguously show the coexistence of two phases , namely a @xmath0-rich phase dominated by a massive @xmath0 cluster that corresponds to the @xmath0-poisoned state and a reactive phase decorated with small @xmath0 islands .
it should be noticed that such a nice coexistence picture is only available using the @xmath49 ensemble since coexistence configurations are not accessible using the standard ensemble .
it should also be noted that the existence of hysteretic effects hinders the location of the coexistence point using the @xmath49 ensemble method .
in fact , in the case of hysteresis in thermodynamic equilibrium the chemical potential at coexistence can be obtained after a proper thermodynamic integration of the growing and decreasing branches of the hysteresis loop @xcite . for nonequilibrium systems like the zgb model where no energetic interactions are considered , the standard methods of equilibrium thermodynamics are not useful . in order to overcome this shortcoming a method based on
the spontaneous creation algorithm already used to study different systems @xcite , has been proposed @xcite .
the method implies the study of the stability of the hysteresis branches upon the application of a small perturbation .
this can be carried out by introducing a negligible small @xmath0 desorption probability ( @xmath195 ) to the zgb model .
it is well known that the first - order nature of the ipt of the zgb model remains if @xmath196 @xcite .
it has been found that taking @xmath197 both branches of the hysteresis loop collapse into a single one that can be identified as the coexistence point given by @xmath198 @xcite .
this figure is close to the middle of the hysteresis loop located close to @xmath199 , where the error bars cover both branches of the loop .
the value @xmath200 reported by brosilow @xcite , which is remarkably close to this figure , has been obtained using the @xmath49 ensemble but neglecting both finite - size and hysteretic effects .
regrettably , the size of the lattice used in reference @xcite was not reported , and therefore additional comparisons can not be performed . in the seminal paper of ziff @xcite , an estimation of the coexistence point
is performed studying the stability of the coexistence phase .
this analysis gives @xmath201 , which is also in good agreement with other independent estimates .
also , evans @xcite have reported @xmath202 based on the analysis of epidemic simulations .
the value reported by meakin @xcite , @xmath203 , seems to be influenced by metastabilities due to the large lattices used for the standard simulation method .
therefore , that figure is a bit larger and may correspond to a value close to the spinodal point .
surprisingly , better values , e.g. @xmath204 @xcite , can be obtained using very small lattices and the standard algorithm since in such samples metastabilities are short - lived .
very recently , hysteresis phenomena have been studied on the basis of a modified version of the zgb model @xcite .
in fact , it is assumed that the surface of the catalyst has two kinds of inhomogeneities or defects . in type-1 defects , which are randomly distributed on the sample with probability @xmath205
, the desorption of adsorbed @xmath0 proceeds with probability @xmath206 . in type-2 inhomogeneities , which are randomly distributed with probability @xmath207
, the adsorption of oxygen molecules is inhibited and the desorption probability of @xmath0 is given by @xmath208 .
furthermore , @xmath209 , where @xmath210 is the desorption probability of @xmath0 on unperturbed lattice sites . also , the diffusion of @xmath0 species is considered with probability @xmath211 , while the probability of other events such as adsorption , desorption and reaction is given by @xmath212 . in order to study hyteretic effects the @xmath0 pressure
is varied at a constant rate in closed cycles .
it is interesting to discus the mechanisms , which are originated by the definition of the model , which may lead to hysteretic effects .
in fact , the low desorption probability of @xmath0 prevents the occurrence of a @xmath0 absorbing state and , consequently , the abrupt transition from the high - reactivity state to the low - reactivity regime is reversible .
furthermore , the blocking of lattice sites for oxygen adsorption also prevents the formation of the oxygen - poisoned state and the second - order ipt becomes reversible and takes place between an oxygen - rich low - reactivity regime and a high - reactivity state .
since escaping from both low - reactivity regimes is quite difficult , the occurrence of hysteresis in dynamic measurements can be anticipated .
the reported results obtained by means of simulations @xcite ( see figure 7 ) are in qualitative agreement with the experimental findings ( figure 3 ) .
figure 7(a ) corresponds to the evolution of oxygen coverage .
@xmath22 decreases smoothly from the oxygen - rich state when the @xmath0 pressure is raised .
this behaviour resembles the case of the zgb model ( figure 1 ) .
however , the abrupt increase in @xmath22 observed when @xmath16 is lowered is due to the fact that the surface is mostly covered by @xmath0 species ( see figure 7(b ) ) , which have low desorption probability .
in fact , oxygen can only be adsorbed on a nearest - neighbor pair of vacant sites that may become available with ( roughly ) a low probability of the order of @xmath213 . on the other hand ,
the growing branch of @xmath21 ( figure 7(b ) ) is flat and that coverage remains close to zero , for @xmath214 .
subsequently , it exhibits an abrupt increase close to @xmath215 , as already expected from the behaviour of the zgb model ( figure 1 ) .
the high coverage region of the decreasing branch is due to the low desorption probability of @xmath0 while the subsequent sudden drop is consistent with the abrupt increase in @xmath22 ( figure 7(b ) ) .
it should be noticed that the experimental measurement of the photocurrent does not allow one to distinguish @xmath21 from @xmath22 @xcite , so figure 3(a ) should be compared with a composition of both figures 7(a ) and 7(b ) .
anyway , the experimental findings are nicely ( qualitatively ) reproduced by the simulations .
finally , the behaviour of the rate of @xmath19 production ( figure 7(c ) is also in excellent agreement with the experimental data shown in figure 3(b ) . 0.5 true cm as in the experiments @xcite ( see figure 3 )
, the numerical simulation results shown in figure 7 also allow one to estimate the values of the @xmath0 partial pressure where the transitions from the monostable states a and b to the bistable state , @xmath216 and @xmath217 , respectively take place .
therefore , the width of the hysteresis loop is given by @xmath218 .
the dependence of @xmath219 on the scan rate of the @xmath0 pressure ( @xmath220 ) has also been measured experimentally and by means of simulations , as shown in figures 8(a ) and 8(b ) , respectively .
again , the numerical results are in remarkable qualitative agreement with the experiments . neglecting surface defects and based on mean - field calculations zhdanov and kasemo
@xcite have stated that the two branches of the hysteresis loop should fall together to the equistability point , provided there occurs a sufficiently rapid nucleation and growth of islands .
this result is expected to be valid on the limit @xmath221 where one also should observe @xmath222 . as shown in figure 8
, the transition points @xmath217 and @xmath216 approach each other and @xmath219 shrinks with decreasing @xmath223 .
conclusive evidence on the validity of the mean - field prediction can not be found due to either experimental or numerical limitations to achieve a vanishing small scanning rate .
however , it has been suggested that @xmath224 should be finite due to the presence of surface defects @xcite , which is neglected in the mean - field treatment @xcite . comparing the numerical results obtained applying the cc method to the zgb model @xcite with those of the study performed by hua @xcite , one has to consider that not only both models are different , but also the data corresponding to the cc ensemble were obtained after a long - time stabilization period and consequently exhibit smaller hysteretic effects , in contrast to the dynamic measurements where the relaxation of the system towards stationary states is not allowed @xcite . as anticipated above a truly @xmath38poisoned state
can not be achieved in the experiments due to the nonvanishing @xmath38desorption probability ( @xmath225 ) . according to the theory of thermal desorption and the experiments , @xmath225 depends on the temperature of the catalysts and the energetic interactions with neighboring adsorbed species through an arrhenius factor @xcite .
therefore , the magnitude of the abrupt drop in the reaction rate ( see figure 2 ) decreases upon increasing @xmath226 , as shown in figure 9 for the case of the catalytic oxidation of @xmath0 on @xmath37 @xcite .
furthermore , on increasing @xmath226 the sharp peak of the reaction rate becomes rounded and for high enough temperature , e.g. for @xmath227 in figure 8 , the signature of the first - order transition vanishes .
the influence of @xmath0 desorption on the phase diagram of the zgb model has also been studied by means of monte carlo simulations @xcite .
the simplest approach is just to introduce an additional parameter to the zgb model , given by the desorption probability @xmath225 .
as expected , the second - order ipt of the models is not influenced by @xmath0 desorption @xcite .
however , the first - order ipt actually disappears because due to the finite value of @xmath225 the system can no longer achieve a truly @xmath38poisoned state .
however , the first - order nature of the transition remains for very low desorption probabilities , as shown in figure 10 , for @xmath228 @xcite . on increasing @xmath225 ,
the peak of the rate of @xmath19 production becomes shifted and rounded in qualitative agreement with the experiments ( figure 9 ) .
1.0 true cm another useful approach to the study of first - order ipt s is to apply the em as described in section .
early epidemic studies of the first - order ipt of the zgb model have been performed by evans and miesch @xcite .
epidemic simulations were started with the surface of the catalysts fully covered by @xmath0 species , except for an empty patch placed at the center of the sample .
the time dependence of the number of empty sites ( @xmath73 ) and the survival probability of the patches ( @xmath74 ) were analyzed in terms of the conventional scaling relationships given by equations ( [ nu ] ) and ( [ so ] ) .
an interesting feature observed in these earlier simulations was the monotonic decrease of @xmath73 , which can be fitted with an exponent @xmath229 @xcite .
this result is in marked contrast with the behaviour expected for second - order ipts in the dp universality class , where equation ( [ nu ] ) holds with a positive exponent such as @xmath230 with @xmath231 @xcite in two dimensions .
furthermore , it has been observed that empty patches have a extremely low survival probability @xcite and the data can be fitted using equation ( [ so ] ) with an exponent @xmath232 , i.e a figure much larger than the exponent expected for dp given by @xmath233 @xcite .
wide experience gained studying * reversible * first - order critical phenomena shows that in this kind of transitions the correlations are short - ranged @xcite .
therefore , the reported power - law decays of @xmath73 and @xmath74 @xcite are certainly intriguing .
however , recent extensive numerical simulations performed averaging results over @xmath234 different epidemic runs have changed this scenario @xcite , as shown in figure 11 . in fact
, data taken for @xmath166 , @xmath180 , and @xmath167 show pronounced curvature with a clear cut - off , departing from a power - law behaviour as described by equation ( [ nu ] ) .
so , it has been concluded that the occurrence of power law ( scale invariance ) in the first - order dynamic critical behaviour of the zgb model can safely be ruled out @xcite . on the other hand , for @xmath235 , log - log plots of @xmath73 versus @xmath75 exhibit pseudo power - law behaviour over many decades ( @xmath236 , as shown in figure 11 .
the effective exponent describing the early time behaviour of @xmath73 is @xmath237 , in agreement with the result reported by evans @xcite .
however , after a long time , few successful epidemics prevail and the number of empty sites suddenly grows as @xmath238 , indicating a spatially homogeneous spreading .
the results shown in figure 11 suggest that instead of the power - law behaviour characteristic of second - order transitions ( equation ( [ nu ] ) ) , the epidemic behaviour close to first - order ipt s could be described by means of a modified ansatz involving a short - time power - law behaviour followed by a long - time exponential decay , as given by equation ( [ anz ] ) @xcite .
the inset of figure 11 shows a test of equation ( [ anz ] ) , namely a semilogarithmic plot of @xmath239 versus @xmath240 , where @xmath241 has been assumed .
the scattering of points for long times is simply due to the reduction of statistics as a consequence of the low survival probability of the initial epidemic patches . summing up ,
epidemic studies of the zgb model close to and at coexistence show a pseudo power - law behaviour for short times ( @xmath242 ) that crosses over to an asymptotic exponential decay for longer times .
consequently , the absence of scale invariance in the first - order ipt of the zgb model places this kind of transition on the same footing its their reversible counterparts .
another interesting scenario for the study of bistable behaviour close to first - order ipt s , observed in catalyzed reactions , is to investigate the properties of interfaces generated during the propagation of reaction fronts .
in fact , key experiments have underlined the importance of front propagation for pattern formation in bistable systems , including the formation of labyrinthine patterns @xcite , self - replicating spots @xcite , target patterns and spiral waves @xcite , stationary concentration patterns ( ` turing structures ' ) @xcite , etc .
furthermore , recent experimental studies of catalytic surface reactions have confirmed the existence of a wealth of phenomena related to pattern formation upon front propagation @xcite .
the basic requirement for the observation of front propagation is a process involving an unstable phase , which could be displaced by a stable one , leading to the formation of an interface where most reaction events take place .
this interesting situation is observed close to first - order ipts , as in the case of the zgb model ( figure 1 ) .
in fact , just at the transition point @xmath35 one has a discontinuity in @xmath243 that corresponds to the coexistence between a reactive state with small @xmath0 clusters and a @xmath0-rich phase , which likely is a large @xmath0-cluster , as suggested by simulations performed using the cc ensemble ( see the snapshots of figures 6(b)-(e ) ) . between @xmath35 and the upper - spinodal point @xmath244
, the reactive state is unstable and it is displaced by the @xmath0-rich phase . on the contrary , between the lower spinodal point @xmath245 and @xmath35 the reactive state displaces the @xmath0-rich phase .
this latter case has been studied by evans and ray @xcite , who have reported that the reactive regime displaces the @xmath0-poisoned state , resulting in a propagation velocity @xmath246 normal to the interface .
it has been proposed that @xmath247 must vanish as @xmath248 @xcite , where both states become equistable , so one has @xmath249 with @xmath250 .
the limit of high diffusivity of the reactants can be well described by mean - field reaction - diffusion equations , which give @xmath251 @xcite .
it is interesting to notice that if diffusion is restricted or even suppressed , simulation results give values of @xmath252 that are also very close to unity , suggesting that this exponent is independent of the surface diffusivity of the reactants @xcite . for an evolving interface
, there is a clear distinction between the propagation direction and that perpendicular to it .
so it may not be surprising that scaling is different along these two directions .
therefore , an interface lacks self - similarity but , instead , can be regarded as a self - affine object @xcite .
based on general scaling arguments it can be shown that the stochastic evolution of a driven interface along a strip of width @xmath158 is characterized by long - wavelength fluctuations @xmath253 that have the following time- and finite - size - behaviour @xcite @xmath254 where @xmath255 for @xmath256 and @xmath257 for @xmath258 , with @xmath259 .
so , the dynamic behaviour of the interface can be described in terms of the exponents @xmath108 and @xmath260 , which are the roughness and growth exponents , respectively .
thus , for an infinite system @xmath261 , one has @xmath262 , as @xmath263 .
note that @xmath264 is also known as the interface width .
it is reasonable to expect that the scaling behaviour should still hold after coarse - graining and passing to the continuous limit .
in fact , the dynamics of an interface between two phases , one of which is growing into the other , is believed to be correctly described by simple nonlinear langevin type equations , such as equation ( [ kkppzz ] ) proposed by kardar , parisi and zhang ( kpz ) @xcite , the edward - wilkinson ( ew ) equation @xcite , and others @xcite .
as in the case of second - order phase transitions , taking into account the values of the dynamic exponents , evolving interfaces can be grouped in sets of few universality classes , such that interfaces characterized by the same exponents belong to the same universality class . among others ,
kpz and ew universality classes are the most frequently found in both experiments and models @xcite , including electrochemical deposition , polycrystalline thin - film growth , fire - front propagation , etc . @xcite . pointing again our attention to the simulation results of evans and ray @xcite ,
they have reported that the propagation of the reaction interface , close to @xmath35 , can be described in terms of dynamic scaling arguments @xcite , with @xmath265 , i.e. , a figure close to the kpz value ( @xmath266 in @xmath267 dimensions ) .
very recently , chvez @xcite studied the dynamics of front propagation in the catalytic oxidation of co on @xmath268 by means of a cellular automaton simulation .
it is found that the dynamic scaling exponents of the interface are well described by equation ( [ fv ] ) with @xmath109 and @xmath266 .
it is also reported that , in the absence of surface diffusion , the interface dynamics exhibits kpz behaviour @xcite .
based on a variant of the zgb model , goodman @xcite have studied the propagation of concentration waves .
they reported the observation of trigger waves within the bistable regime of the process , i.e. , close to the first - order ipt .
in fact , within this regime one has the coexistence of a stable state with a metastable one . at the boundary between the two
, the stable state will displace the metastable one and the boundary will move , so this process leads to the propagation of concentration fronts ( trigger waves ) .
goodman @xcite found that the velocity of the @xmath0 front depends on the diffusion rate @xmath269 of @xmath0 species ( diffusion of oxygen is neglected ) and @xmath270 .
the velocity of the front vanishes on approaching the poisoning transition at @xmath271 ( note that the transition point now depends on @xmath269 ) , according to equation ( [ velo ] ) , with @xmath272 , in agreement with the results of evans @xcite .
while front propagation during the catalytic oxidation of @xmath0 on platinum surfaces has been observed in numerous experiments @xcite , the quantitative analysis of data is somewhat restricted by the fact that the fronts are neither flat nor uniformly curved , eventually several of them nucleate almost at the same time and , in most cases , the occurrence of strong interactions between fronts ( ` interference of chemical waves ' ) makes clean interpretations quite difficult . in order to overcome these shortcomings ,
haas @xcite have studied the propagation of reaction fronts on narrow channels with typical widths of 7 , 14 and 28 @xmath273 .
the main advantage of these controlled quasi - one - dimensional geometries is that frequently only a single front propagates along the channel , thus avoiding front interactions .
additionally , the narrowness of the channels and the absence of undesired flux at the boundaries lead to practically planar fronts . using this experimental setup ,
the front velocity in different channels can be measured as a function of the partial pressure of @xmath0 , keeping the temperature and the oxygen partial pressure constant , as shown in figure 12(a ) . at low @xmath16 values
only oxygen fronts are observed .
furthermore , their velocity decreases when @xmath16 is increased , reaching a minimum value at a certain critical threshold @xmath274 ( see figure 12(a ) ) .
when @xmath16 is further increased a jump is observed- now the front reverses itself into a @xmath0 front and travels in the opposite direction .
when @xmath16 is lowered from high values , the @xmath0 fronts become slower and hysteresis is observed ( see the coexistence between oxygen and @xmath0 fronts in figure 12(a ) for @xmath275 ) .
finally , at @xmath276 another jump is observed- under these conditions @xmath0 fronts can no longer persist below a quite low ( but nonzero velocity ) and they reverse themselves into fast oxygen fronts ( figure 12(a ) ) .
many features of the experiment of haas @xcite can be recovered simulating front propagation with the aid of the zgb model on the square lattice with rectangular geometries of sides @xmath277 ( @xmath278 ) @xcite .
thus @xmath158 is the width of the channel and @xmath279 its length .
free boundary conditions were taken along the channel while the opposite ends are assumed to be in contact with oxygen and @xmath0 sources , respectively .
if @xmath18 or @xmath0 species are removed from the ends of the channels ( i.e. , the ` sources ' ) , due to the reaction process , they are immediately replaced .
the propagation of the @xmath0 concentration profile was studied starting with a sample fully covered by oxygen , except for the first and second columns , which are covered by @xmath0 ( the @xmath0 source ) , and left empty , respectively .
the propagation of the oxygen profile was followed using a similar procedure @xcite . under these conditions
one always has two competing interfaces along the channel . in order to make a quantitative description of the propagation
, the concentration profiles of the reactants , @xmath280 and @xmath281 , are measured along the length of the channel @xmath282 in the @xmath279-direction and averaged over each column of lattice sites of length @xmath158 .
then , the moments of the profiles , which in subsequent steps can be used to determine the propagation velocity and the width of the profiles , are also measured .
in fact , the moments of @xmath283th order of the profiles can be evaluated according to @xcite @xmath284}{\sum [ \theta(x+1 ) - \theta(x ) ] } .
\label{moment}\ ] ] thus , using equation ( [ moment ] ) the velocity of propagation can be obtained from the first moment @xmath285 monte carlo simulation results show that the front propagation velocity depends on both @xmath16 and the channel width @xmath158 , as shown in figure 12(b ) .
this figure also shows that the displacement of @xmath0- and @xmath18-poisoned channels by the reactive regime stops at certain ( @xmath158-dependent ) critical values , @xmath286 and @xmath287 , respectively . by means of an extrapolation to the thermodynamic limit it is possible to identify these critical values with the critical points of the zgb model , namely @xmath288 and @xmath289 , respectively .
it is also found that close to @xmath290 , when the propagation of the @xmath18 profile ceases , the velocity of the @xmath0 profile undergoes a sharp change .
this behaviour can be correlated with the first - order ipt between the stationary reactive regime and the @xmath0-poisoned state observed in the zgb model at @xmath35 ( see figure 1 ) .
so far , the main conclusions that can be drawn from figure 12(b ) can be summarized as follows : a ) there are two critical pressures , @xmath287 and @xmath286 , which depend on the width of the channel , at which propagation of one profile or the other stops ; b ) within these critical values , propagating @xmath0 and @xmath18 profiles coexist ; c ) @xmath18 profiles propagate faster than @xmath0 profiles .
all these observations appear in qualitative agreement with the experimental results shown in figure 12(a ) @xcite .
however , the underlying physics is different : in the simulations the displacement of a poisoned phase by the invading reactive phase takes place within a range of pressures where the latter is unstable , while the former is stable . in contrast , the experiment may show the propagation of coexisting phases within a bistable regime @xcite . so far , all those simulations of the zgb model discussed above do not attempt to describe the occurrence of oscillations in the concentration of the reactants and in the rate of production , which are well documented by numerous experiments @xcite .
in fact , it is well known that the catalytic oxidation of @xmath0 on certain @xmath15 surfaces exhibits oscillatory behaviour , within a restricted range of pressures and temperatures , which is associated with adsorbate - induced surface phase transitions @xcite .
since the aim of this paper is to describe the irreversible critical behaviour of the reaction , the oscillatory behaviour will not be further discussed .
therefore , the interested reader is addressed to recent developments involving the study of numerous lattice - gas models aimed to explain the oscillations observed experimentally @xcite . since the zgb lattice gas reaction model is an oversimplified approach to the actual processes involved in the catalytic oxidation of co ,
several attempts have been made in order to give a more realistic description .
some of the additional mechanisms and modifications added to the original model are the following : ( i ) the inclusion of @xmath0 desorption @xcite that causes the first - order ipt to become reversible and slightly rounded , in qualitative agreement with experiments ( figures 9 and 10 ) .
( ii ) energetic interactions between reactants adsorbed on the catalyst surface have been considered by various authors @xcite . in general ,
due to these interactions the ipt s become shifted , rounded and occasionally they are no longer observed @xcite .
( iii ) studies on the influence of the fractal nature of the catalyst are motivated by the fact that the surface of most solids at the molecular level must be considered as a microscopic fractal , such as the case of cataliysts made by tiny fractal matallic cluster dispersed in a fractal support or in a discontinuous thin metal films .
the fractal surfaces have been modeled by means of random fractals , such as percolating clusters , @xcite , diffusion limited aggregates @xcite and also deterministic fractals , such as sierpinsky carpets , @xcite , etc . @xcite .
one of the main findings of all these studies is that the first - order ipt becomes of second - order for dimensions @xmath291 . since in @xmath292 dimensions the zgb model
does not exhibit a reaction window , one may expect the existence of a ` critical ' lower fractal dimension capable a sustaining a reactive regime .
this kind of study , of theoretical interest in the field of critical phenomena , remains to be addressed .
( iv ) also , different kinds of adsorption mechanisms , such as hot - dimer adsorption @xcite , local versus random adsorption @xcite , nonthermal mechanisms involving the precursor adsorption and diffusion of @xmath0 molecules @xcite , the presence of subsurface oxygen @xcite , etc . have been investigated .
( v ) the influence of surface diffusion has also been addressed using different approaches @xcite .
particularly interesting is the hybrid lattice - gas mean - field treatment developed by evans @xcite for the study of surface reactions with coexisting immobile and highly mobile reactants .
( vi ) considering the eley - rideal mechanism @xcite as an additional step of the set of equations ( [ adco]-[reac ] ) , namely including the following path @xmath293 poisoning of the surface by complete occupation by @xmath18 species is no longer possible preventing the observation of the second - order ipt .
( vii ) the influencia of surface defects , which has also been studied , merits a more detailed discussion because , in recent years , considerable attention has been drawn to studies of surface reactions on substrates that include defects or some degrees of geometric heterogeneity , which is not described by any fractal structure , as in the case of item iii ) .
interest in these types of substrates is based on the fact that the experiments have shown that inert species capable to block adsorption sites , such a sulfur , deposit on the catalyst surface during the exhaust of the combustion gases .
also crystal defects that are formed during the production of the catalyst result in blocked sites .
other inhomogeneities consist of crystallographic perturbations but may equally well involve active foreign surface atoms .
recently lorenz @xcite have performed monte carlo simulations using three different types of defective sites .
site-1 , that adsorbed neither @xmath18 nor @xmath0 and site-2 ( site-3 ) that adsorbed @xmath18 ( @xmath0 ) but no @xmath0 ( @xmath18 ) .
they found that @xmath0 islands form around each defect near @xmath35 ( @xmath294 ) .
the average density of @xmath0 decays as a power - law of the radial distance from the defect ( @xmath295 , @xmath296 ) , and the average cluster size also obeys a power - law with the distance to spinodal point ( @xmath297 ) with exponent @xmath298 .
when defects are randomly distributed , with density @xmath299 , @xmath35 decreases linearly according to @xmath300 .
this model has also been investigated in the site and pair mean - field approximations @xcite .
the pair approximation exhibits the same behaviour that the monte carlo simulation .
the size of the reactive windows decreases with @xmath299 and the abrupt transition at @xmath35 becomes continuous ( the same behaviour have been reported in a related model @xcite ) . however , unlike the analytical results , in the monte carlo simulation there is a critical concentration above which the transition at @xmath35 becomes continuous ( @xmath301 in agreement with previous results @xcite ) . in conclusion , various models shown that the presence of defects on the catalytic surface causes the @xmath0 poisoning transition to occur at lower @xmath16 values than on homogeneous surfaces .
also , beyond some critical concentration of defects , the first - order ipt of the zgb model becomes second - order .
the overall effect of inert sites is to reduce the production of @xmath302 .
furthermore , these findings provide an alternative explanation for the absence of a second - order ipt into a @xmath18-poisoned state observed in the experiments of @xmath0 oxidation ( see figure 2 ) . the catalytic reduction of @xmath303 with various agents , including @xmath0 , @xmath7 , @xmath304 , hydrocarbons , etc .
, has been extensively studied on @xmath15 and @xmath305 surfaces @xcite , which are the noble metals used in automotive catalytic converters , due to the key role played by @xmath1 emission in air pollution @xcite . aside from the practical importance
, the catalytic reduction of @xmath303 also exhibits a rich variety of dynamic phenomena including multistability and oscillatory behaviour @xcite . within this context ,
the catalytic reaction between @xmath303 and @xmath0 is the subject of this section .
the archetypal model used in most simulations has early been proposed by yaldran and khan ( yk ) @xcite . as in the case of the zgb model @xcite ,
the yk model is also a lattice gas reaction system based on the langmuir - hinshelwood mechanism .
the reaction steps are as follows @xmath306 @xmath307 @xmath308 @xmath309 where @xmath12 represents an unoccupied site on the catalyst surface , @xmath310 represents a nearest neighbor ( nn ) pair of such sites , @xmath311 indicates a molecule in the gas phase and @xmath312 indicates an species adsorbed on the catalyst .
the reactions given by equations ( [ reco ] ) and ( [ ren ] ) are assumed to be instantaneous ( infinity reaction rate limit ) while the limiting steps are the adsorption events given by equations ( [ noad ] ) and ( [ coad ] ) .
the yk model is similar to the zgb model for the @xmath148 reaction , except that the @xmath6 is replaced by @xmath303 , and nn @xmath313 atoms , as well as nn @xmath190 pairs , react . for further details on the yk model
see @xcite .
early simulations of the yk model have shown @xcite that a reactive window is observed on the hexagonal lattice while such kind of window is absent on the square lattice @xcite , pointing out the relevance of the coordination number for the reactivity .
therefore , we will first discuss monte carlo simulations of the yk model performed on the hexagonal ( triangular ) lattice .
subsequently , results obtained for variants of the yk model that also exhibit reaction windows on the square lattice will be presented .
the simulation procedure in the standard ensemble is as follows : let @xmath314 and @xmath16 be the relative impingement rates for @xmath303 and @xmath0 , respectively , which are taken to be proportional to their partial pressures in the gas phase
. taking @xmath315 , such normalization implies that the yk model has a single parameter that is usually taken to be @xmath16 . @xmath0 and
@xmath303 adsorption events are selected at random with probabilities @xmath16 and @xmath316 , respectively .
subsequently , an empty site of the lattice is also selected at random .
if the selected species is @xmath0 , the adsorption on the empty site occurs according to equation ( [ coad ] ) .
if the selected molecule is @xmath303 , a nn site of the previously selected one is also chosen at random , and if such site is empty the adsorption event takes place according to equation ( [ noad ] ) .
of course , if the nn chosen site is occupied the adsorption trial is rejected .
after each successful adsorption event all nn sites of the adsorbed species are checked at random for the occurrence of the reaction events described by equations ( [ reco ] ) and ( [ ren ] ) . during the simulations , the coverages with @xmath0 , @xmath18 and @xmath313 ( @xmath21 , @xmath22 and @xmath317 , respectively ) as well as the rate of production of @xmath19 and @xmath5 ( @xmath20 , @xmath318 , respectively ) are measured .
the phase diagram of the yk model , shown in figure 13 , is similar to that of the zgb model @xcite shown in figure 1 .
in fact , in both cases second- and first - order ipt s are observed .
however , in contrast to the zgb model where the absorbing ( poisoned ) states are unique , in the case of the yk model such states are mixtures of @xmath319 and @xmath320 as follows from the observation of the left and right sides of the phase diagram , respectively ( figure 13(a ) ) .
the ipt observed close to @xmath321 @xcite is continuous and therefore of second - order ( see figure 13 ) .
more interesting , an abrupt first - order ipt is also observed close to @xmath322 ( figure 1(a ) and ( b ) ) @xcite .
hysteretic effects close to the first - order ipt of the yk model have been investigated using the cc ensemble @xcite ( see ) . for small lattices ( @xmath323 )
the relaxation time is quite short , so that hysteretic effects are absent .
this result is in agreement with similar measurements of the zgb model ( see figure 4(a ) ) . on increasing the lattice size ,
hysteretic effects can be observed even for @xmath324 and they can unambiguously be identified for @xmath325 , as shown in figure 14(a ) . a vertical region located at the center of the loop and slightly above @xmath326 , as well as the upper spinodal point @xmath327 , can easily be observed in figure 14 .
furthermore , it is found that while the location of @xmath327 is shifted systematically toward lower values when @xmath158 is increased , the location of the vertical region ( close to the center of the loops ) remains almost fixed very close to @xmath328 @xcite .
using lattices of size @xmath329 , the hysteretic effects are quite evident ( see figure 14(b ) ) and also , the growing and decreasing branches of the loops are almost vertical .
also , the location of these branches depends on the lattice size , as follows from the comparison of figures 14(a ) and ( b ) . versus @xmath330 obtained using the cc ensemble and taking : ( a ) @xmath325 and ( b ) @xmath172 .
the arrows indicate the growing branch ( gb ) , the decreasing branch ( db ) , the vertical region(vr ) and the upper spinodal point ( @xmath165 ) .
more details in the text.,width=226 ] versus @xmath330 obtained using the cc ensemble and taking : ( a ) @xmath325 and ( b ) @xmath172 .
the arrows indicate the growing branch ( gb ) , the decreasing branch ( db ) , the vertical region(vr ) and the upper spinodal point ( @xmath165 ) .
more details in the text.,width=226 ] a more quantitative analysis on the behaviour of @xmath16 corresponding to the different branches and the vertical region has also been reported @xcite . versus @xmath169
measured in the growing branch ( @xmath331 ) , decreasing branch ( @xmath332 ) , and the vertical region ( @xmath333 ) .
the straight lines correspond to the best fits of the data that extrapolate to @xmath334 .
( b ) plots of @xmath165 versus @xmath169 .
the straight line corresponds to the best fit of the data that extrapolates to @xmath335 .
more details in the text.,width=226 ] versus @xmath169 measured in the growing branch ( @xmath331 ) , decreasing branch ( @xmath332 ) , and the vertical region ( @xmath333 ) .
the straight lines correspond to the best fits of the data that extrapolate to @xmath334 .
( b ) plots of @xmath165 versus @xmath169 .
the straight line corresponds to the best fit of the data that extrapolates to @xmath335 .
more details in the text.,width=226 ] figure 15(a ) shows the dependence of the location of the growing branch and the decreasing branch ( @xmath336 and @xmath337 , respectively ) on the inverse of the lattice size .
the @xmath158 dependence of @xmath16 at the vertical region ( @xmath338 ) is also shown for the sake of comparison .
it has been found that the location of all relevant points , namely @xmath339 with @xmath340 and @xmath341 , depends on the curvature radius ( @xmath342 ) of the interface of the massive @xmath0 cluster in contact with the reactive region .
such dependence can be written as follows @xmath343 where @xmath344 is the location of the point under consideration after proper extrapolation to the thermodynamic limit and @xmath345 is an @xmath342-dependent function .
for the vertical region one has @xmath346 and @xmath338 is almost independent of @xmath158 , so @xmath347 , as shown in figure 15(a ) .
in contrast , for the db and the gb , @xmath342 is finite and of the order of @xmath348 and @xmath349 , respectively .
so , one has @xmath350 while @xmath351 , in agreement with the results shown in figure 15(a ) . the extrapolated points are @xmath352 , @xmath353 and @xmath354 also
, @xmath355 and @xmath356 have been @xcite . on the basis of these results ,
@xmath357 has been identified as the coexistence point @xmath358 in excellent agreement with an independent measurement , @xmath359 , reported by brosilow and ziff @xcite .
this result is in contrast with measurements performed with the zgb model .
in fact , for the zgb systems the vertical region is not observed while the locations of the growing and decreasing branches are almost independent of the lattice size ( see figure 4 ) .
the explanation of the difference observed comparing both models , which may be due to the different behaviour of the interface of the massive @xmath0 cluster , is an interesting open question .
it has also been reported that the location of the upper spinodal point depends on the lattice size , as shown in figure 15(b ) .
this dependence of @xmath360 is due to local fluctuations in the @xmath0 coverage that take place during the nucleation of the critical cluster @xcite .
the extrapolation of the data shown in figure 15(b ) gives @xmath361 .
furthermore , the coverage at this point is @xmath362 .
these results point out that in the thermodynamic limit the spinodal point is very close to coexistence , i.e. , @xmath363 . for the sake of comparison
it is worth mentioning that for the zgb model one has @xmath364 ( see ) .
further insight into the first - order ipt of the yk model can be gained performing epidemic studies . however , in this case it is necessary to account for the fact that the poisoned ( absorbing ) state above coexistence is nonunique , since it is due to a mixture of @xmath0 and @xmath313 atoms with coverage @xmath365 and @xmath366 , as shown in figure 13(a ) .
so , the starting configuration has to be obtained running the actual dynamics of the system slightly above coexistence until ` natural ' absorbing states suitable for the studies are generated .
figure 16 shows results obtained performing epidemic simulations for various values of @xmath16 including @xmath367 , @xmath244 , @xmath167 , @xmath166 as well as a value close to coexistence but slightly inside the active region , namely @xmath368 .
= 6.0 cm from figure 16 , it becomes evident that the method is quite sensitive to tiny changes of @xmath16 .
the obtained curves are fitted by equation ( [ anz ] ) with @xmath369 , indicating a markedly low survivability of the epidemic patches as compared with the zgb model that gives @xmath370 , as already discussed in .
the main finding obtained using epidemic studies is that the occurrence of a power - law scaling behaviour close to coexistence can unambiguously be ruled out .
this result is in qualitative agreement with data of the zgb model , see .
all these observations are also in agreement with the experience gained studying first - order reversible phase transitions where it is well established that correlations are short ranged , preventing the emergence of scale invariance .
it should be mentioned that several mean - field theories of the yk model have been proposed @xcite . truncating the hierarchy of equations governing the cluster probabilities at the @xmath371site level , a reasonable estimate of the coexistence point given by @xmath372
is obtained @xcite on the triangular lattice .
however , this approach fails to predict the second - order ipt observed in the simulations ( see figure 13 ) @xcite .
also , kortlke @xcite have derived an accurate prediction of the second - order critical point at @xmath373 using a two - site cluster approximation .
the prediction of this approach for the coexistence point is less satisfactory , namely @xmath374 . on the other hand ,
very recently an elaborated mean - field theory of the yk model has been developed up to the pair - approximation level @xcite that yields a very accurate estimation of the coexistence point , namely @xmath375 . as already mentioned above , the behaviour of the yk model on the square lattice is radically different than that observed on the triangular lattice .
in fact , in the former * no * reactive stationary state has been observed @xcite , as shown in figure 17(a ) . .
@xmath22 ( empty triangles ) , @xmath317 ( empty squares ) , and @xmath21 ( solid circles ) .
( a ) results obtained neglecting @xmath313 diffusion showing the absence of a reaction window such that the catalyst always remains poisoned by mixtures of adsorbed species .
( b ) results obtained considering @xmath313 diffusion . in this case
the yk model exhibits a reaction window .
adapted from reference @xcite .
( c ) results obtained considering the influence of the eley - rideal ( er ) mechanism and the diffusion on @xmath313 . adapted from reference @xcite .
more details in the text.,width=226 ] .
@xmath22 ( empty triangles ) , @xmath317 ( empty squares ) , and @xmath21 ( solid circles ) .
( a ) results obtained neglecting @xmath313 diffusion showing the absence of a reaction window such that the catalyst always remains poisoned by mixtures of adsorbed species .
( b ) results obtained considering @xmath313 diffusion . in this case
the yk model exhibits a reaction window .
adapted from reference @xcite .
( c ) results obtained considering the influence of the eley - rideal ( er ) mechanism and the diffusion on @xmath313 . adapted from reference @xcite .
more details in the text.,width=226 ] however , simulation results have shown that diffusion of @xmath313 ( but not of @xmath18 , or @xmath0 ) restores the possibility of a reactive state @xcite , as shown in figure 17(b ) .
in fact , in this case a second - order ipt is observed close to @xmath376 , while a first - order ipt is found at @xmath377 @xcite .
also , meng @xcite have shown that by adding a new reaction channel to yk model ( equations ( [ noad]-[ren ] ) ) , such that @xmath378 the reactivity of the system becomes enhanced and consequently a reaction window is observed on the square lattice .
this window exhibits a second - order ipt close to @xmath379 and a first - order ipt close to @xmath380 .
this behaviour is reminiscent of that observed modeling the zgb model , as discussed above . on the other hand , assuming that the dissociation of @xmath303 given by equation ( [ noad ] ) is preceded by a molecular adsorption on a singe site , namely @xmath381 and @xmath382 the yk model also exhibits a reaction window in the square lattice provided that both @xmath303 and @xmath0 desorption are considered @xcite .
very recently , khan @xcite have studied the influence of the eley - rideal ( er ) mechanism ( reaction of @xmath0 molecules with already chemisorbed oxygen atoms to produce desorbing @xmath19 ) on the yk model on the square lattice . in the absence of @xmath313 diffusion ,
the added er mechanism causes the onset of a reactive regime at extremely low @xmath0 pressures , i.e. , for @xmath383 . however , considering the diffusion of @xmath313 species , the window becomes considerably wider and the reactive regime is observed up to @xmath384 where a first - order itp is found @xcite , as shown in figure 17(c ) .
this finding suggests that the incorporation of the er mechanisms does not affect the first - order ipt ( see figure 17(b ) ) .
in contrast , the second - order ipt is no longer observed as shown in figure 17(c ) . as in the case of the zgb model ,
the bistable behaviour of the yk model close to coexistence provides the conditions for the displacement of reactive fronts or chemical waves . within this context , tammaro and evans @xcite
have studied the reactive removal of unstable mixed @xmath385 layers adsorbed on the lattice .
furthermore , in order to account for the diffusion of the reactants , the hopping of all adsorbed species ( except for @xmath18 atoms whose mobility is negligible ) has been considered .
simulations are started with the surface fully covered by a mixture @xmath385 .
this mixture is unstable since the vacation of a single site may produce the dissociation of @xmath303 ( equation ( [ noaddiso ] ) ) and its subsequent reaction with @xmath0 followed by desorption of the products and the generation of empty sites capable of triggering the autocatalytic reaction . due to the high mobility of most adsorbed species , initially an exponential increase in the number of highly dispersed vacancies is observed .
thereafter , a reaction front forms and propagates across the surface at constant velocity @xcite .
it is also interesting to remark that all simulation results are confirmed by an elaborated mean - field treatment of chemical diffusion on mixed layers , incorporating its coverage - dependent and tensorial nature , both of these features reflecting the interference of chemical diffusion of adsorbed species on surface by coadsorbed species @xcite .
in addition to the zgb and the yk models , numerous lattice gas reaction models have also been proposed attempting to describe catalyzed reaction processes of practical and academic interest . among others ,
the dimer - dimer ( dd ) surface reaction scheme of the type @xmath386 has been proposed in order to describe the catalytic oxidation of hydrogen @xcite .
monte carlo simulations of the dd model have shown the existence of second - order ipt s and a rich variety of irreversible critical behaviour @xcite .
relevant numerical results have also been qualitatively reproduced by mean - field calculations @xcite . on the other hand ,
the catalytic synthesis of ammonia from hydrogen and nitrogen on iron surfaces ( @xmath387 ) is among the catalyzed reactions of major economical importance .
ever since its discovery and technical realization , the reaction has become the focus of fundamental investigations .
very recently , various lattice gas reaction models have been proposed and studied by means of numerical simulations @xcite .
the existence of ipt s has been observed , as shown in figure 18 for the case of the model proposed by khan and ahmad ( ka ) @xcite . here
the langmuir - hinshelwood reaction mechanism is assumed and the control parameter is the partial pressure of @xmath7 .
as follows from figure 18 , the ka model exhibits a second - order ( first - order ) ipt close to @xmath388 ( @xmath389 ) , resembling the behaviour of both the zgb and the yk models ( see figures 1 and 13 , respectively ) . .
@xmath390 ( open circles ) , @xmath317 ( empty triangle ) @xmath391 ( open squares ) and @xmath392 ( solid circles ) . adapted from reference @xcite.,width=226 ] recently , the simulation of reaction processes involving more than two reactants has received growing attention .
one example of these processes is a multiple - reaction surface reaction model based upon both the zgb and the dd models @xcite .
so , this zgb - dd model may be applied to the oxidation of co in the presence of @xmath393-traces as well as to the oxidation of hydrogen in the presence of co - traces .
interest on this model is due to various reasons .
for example , the oxidation of hydrogen and carbon monoxide plays a key role in the understanding of the process of hydrocarbon oxidation .
in fact , the oxy - hydrogen reaction mechanism contains the chain - branching steps producing o , h and oh - radicals that attack hydrocarbon species . also , co is the primary product of hydrocarbon oxidation and it is converted to carbon dioxide in a subsequent slow secondary reaction .
furthermore , the zgb - dd model exhibits interesting irreversible critical behaviour with nonunique multi - component poisoned states @xcite .
there are also models that are not aimed to describe any specific reaction system but , instead , they are intended to mimic generic reactions .
typical examples are the monomer - monomer model @xcite , the dimer - trimer model @xcite , the monomer - trimer model @xcite , etc .
( see also references @xcite .
on the other hand , in the literature there is a vast variety of lattice gas reaction models following the spirit of the reaction described in the above subsections .
they all exhibit the same type of irreversible critical behaviour at the transition , which is determined by a common feature- the existence of an absorbing or poisoned state , i.e. , a configuration that the system can reach but from where it can not escape anymore , as discussed in detail in section 1.2 .
as already discussed , the essential physics for the occurrence of ipt s is the competition between proliferation and death of a relevant quantity .
so , it is not surprising that a large number of models satisfying this condition , and aimed to describe quite diverse physical situations , have been proposed and studied . some examples of models exhibiting second - order ipt s are , among others ` directed percolation ' as a model for the systematic dripping of a fluid through a lattice with randomly occupied bonds @xcite , the ` contact process ' as a simple lattice model for the spreading of an epidemics @xcite , ` autocatalytic reaction - diffusion models ' aimed to describe the production of some chemical species @xcite , ` the stochastic game of life ' as a model for a society of individuals @xcite , ` forest fire models ' @xcite , ` branching annihilating random walkers ' with odd number of offsprings @xcite , epidemic spreading without immunization @xcite , prey - predator systems @xcite , the domany - kinzel cellular automata @xcite , etc . for an excellent review on this subject see @xcite .
the common feature among all these models is that they exhibit second - order ipt s belonging to the universality class of directed percolation ( dp ) , the langevin equation ( ) being the corresponding field - theoretical representation . the robustness of these dp models with respect to changes in the microscopic dynamic rules is likely their most interesting property .
such robustness has led janssen and grassberger @xcite to propose the so - called dp conjecture , stating that models in the dp universality class must satisfy the following conditions : ( i ) they must undergo a second - order ipt from a fluctuating active state to a unique absorbing state .
( ii ) the transition has to be characterized by a positive single - component order parameter .
( iii ) only short - range processes are allowed in the formulation of the microscopic dynamic rules .
( iv ) the system has neither additional symmetries nor quenched randomness . in spite of the fact that the dp conjecture has not been proved rigorously , there is compelling numerical evidence supporting it @xcite .
so far , dp appears to be the generic universality class for ipt s into absorbing states , having a status similar to their equilibrium counterpart , namely the venerated ising model .
however , despite the successful theoretical description of the dp process , there are still no experiments where the critical behaviour of dp has been observed .
therefore , this is a crucial open problem in the field of ipt s . for further discussions
see for instance @xcite .
the study of irreversible critical behaviour in reaction systems has attracted the attention of many physicists and physical - chemists for more than four decades . on the one hand ,
second - order ipt s are quite appealing since , like the archetypal case of dp , they are observed in simple models in terms of their dynamic rules .
nevertheless , second - order behaviour is highly nontrivial and has not yet been solved exactly , even using minimal models in one dimension .
furthermore , the critical exponents are not yet known exactly .
field - theoretical calculations and numerical simulations have greatly contributed to the understanding of second - order irreversible behaviour .
most systems lie in the universality class of dp , which plays the role of a standard universality class similar to the ising model in equilibrium statistical physics , and the reason for few of the exceptions found are very well understood .
the main challenges in the field are , from the theoretical point of view , the achievement of exact solutions even for simple models and , from the point of view of the experimentalists , the realization of key experiments unambiguously showing dp behaviour .
on the other hand , the scenario for the state of the art in the study of first - order ipt s is quite different .
firstly , there is stimulating experimental evidence of the existence of abrupt ( almost irreversible ) transitions , while hysteretic effects and bistable behaviour resembling first - order like behaviour have been observed in numerous catalyzed reaction experiments .
secondly , one still lacks a theoretical framework capable describing first - order ipt s and theoretical efforts are being addressed to the development of mean - field approaches with different degrees of sophistication .
the achievement of a theoretical framework enabling the treatment of irreversible critical behaviour and the gathering of further experimental evidence , including the accurate measurement of critical exponents , are topics of high priority that will certainly contribute to the development of a general theory of the physics of far - from equilibrium processes .
this work is financially supported by conicet , unlp and anpcyt ( argentina ) .
we are grateful with numerous colleagues for stimulating discussions .
hilderbrand m , kuperman m , wio h , mikhailov a and ertl g 1999 1475 only few exceptions of dp are known so far .
however , they all violate at least one of the essential conditions required for the validity of the dp conjecture , as dicussed in . for some examples
see reference @xcite . | an introductory review on the critical behaviour of some irreversible reaction systems is given .
the study of these systems has attracted great attention during the last decades due to , on the one hand , the rich and complex underlying physics , and on the other hand , their relevance for numerous technological applications in heterogeneous catalysis , corrosion and coating , development of microelectronic devices , etc .. the review is focuses on recent advances in the understanding of irreversible phase transitions ( ipt s ) providing a survey of the theoretical development of the field during the last decade , as well as a detailed discussion of relevant numerical simulations .
the langevin formulation for the treatment of second - order ipt s is discussed .
different monte carlo approaches are also presented in detail and the finite - size scaling analysis of second - order ipt s is described .
special attention is devoted to the description of recent progress in the study of first - order ipt s observed upon catalytic oxidation of carbon monoxide and the reduction of nitrogen monoxide , using lattice gas reaction models .
only brief comments are given on other reactions such as the oxidation of hydrogen , ammonia synthesis , etc .
also , a discussion of relevant experiments is presented and measurement are compared with the numerical results .
furthermore , promising areas for further research and open questions are also addressed .
= 20pt # 1 |
we already have a precise measurement of @xmath0 from @xmath1
. however this measurement does not resolve the 4 fold ambiguity in @xmath2 : @xmath6 we want to measure the angle @xmath2 in many processes to test the sm .
also measurement of both @xmath0 and @xmath3 is clearly desirable to partly resolve the discrete ambiguity in @xmath2 . in this talk
i will review the theory of measuring the ckm angle @xmath4 from various @xmath5 decays to charm final states .
i will concentrate on @xmath7 , @xmath8 .
these are @xmath9 transitions .
i will then discuss @xmath10 ( @xmath11 ) and @xmath12 ( @xmath13 ) .
there are other approaches to measuring @xmath14 that will not be discussed in this paper @xcite .
the decay @xmath7 is a vv decay and there are three amplitudes- @xmath15 , @xmath16 which are cp even and @xmath17 which is cp odd . the time dependent angular distribution allows us to measure both @xmath18 and @xmath19 .
the time independent angular distributions is needed to fix the coefficients of in the time dependent distributions @xcite . however the time independent analysis gives the triple products , @xcite @xmath20 but the coefficient of @xmath19 term depend on the cosine of the phase differences @xmath21 and @xmath22 . to resolve this ambiguity in the coefficient term information
is obtained from the decay @xmath23 .
the @xmath24 system originating from @xmath25 can , in principle , have any integer spin but it is found experimentally that , below 1.3 gev , the @xmath26 and @xmath27 waves dominate . assuming that the strong interactions between the @xmath28 and the @xmath29 are small , one can obtain additional information about the p - wave phase shift by an angular analysis of the decay @xmath25 .
this additional information can then be used to resolve the sign ambiguity in @xmath19 measurement .
the result of such an analysis yields a positive value for @xmath19 @xcite , which combined with that the value of @xmath18 obtained in @xmath1 results in a value of @xmath2 which is consistent with standard model expectations .
this decay can have both non resonant and resonant contributions .
the resonant contributions can go through an intermediate excited @xmath31 making this process sensitive to @xmath19 measurement @xcite .
we define the following amplitudes & & a(b^0(p)d^+*__1(p_+ ) d^-*__2(p_- ) k_s(p_k ) ) , & & a(|b^0(p)d^+*__1(p_+ ) d^-*__2(p_- ) k_s(p_k ) , [ aabar ] where @xmath32 and @xmath33 represent unmixed neutral @xmath5 and @xmath34 and @xmath35 are the polarization indices of the @xmath36 and @xmath37 respectively . the decay @xmath38.,scaledwidth=40.0% ]
the time - dependent amplitudes for an oscillating state @xmath39 which has been tagged as a @xmath32 meson at time @xmath40 is given by a^_1,_2(t ) & = & a^_1,_2 ( 2 ) + i e^-2i + & & |a^_1,_2 ( 2),and the time - dependent amplitude squared summed over polarizations and integrated over the phase space angles is : g_0(s^+,s^- ) & = & |a(s^+,s^-)|^2 + ||a(s^+,s^-)|^2 , + g_c(s^+,s^- ) & = & |a(s^+,s^-)|^2 -||a(s^+,s^-)|^2 , + g_s(s^+,s^- ) & = & 2(e^-2i |a(s^+,s^- ) a^(s^+,s^- ) ) + & = & -2 ( 2 ) ( |aa^ ) + 2(2 ) ( |aa^ ) , + s^+ & = & ( p_+ + p_k)^2,s^-= ( p_- + p_k)^2 , where @xmath41 .
it is convenient to replace @xmath42 where @xmath43 is the @xmath44 energy in the rest frame of the @xmath5 .
the variable @xmath45 is defined as , @xmath46 with @xmath47 being the angle between the momentum of @xmath44 and @xmath36 in a frame where the two @xmath48 are moving back to back along the z- axis .
note that , @xmath49 corresponds to @xmath50 .
the definition of the variable @xmath45.,scaledwidth=40.0% ] carrying out the integration over the phase space variables @xmath45 and @xmath43 one gets the following expressions for the time - dependent total rates for @xmath51 and the cp conjugate process , ( t ) & = & [ i_0 + 2 ( 2)(m t ) i_s1 ] , + ( t ) & = & [ i_0 - 2 ( 2 ) ( m t ) i_s1],where @xmath52 and @xmath53 are the integrated @xmath54 and @xmath55 functions .
one can then extract @xmath56 from the rate asymmetry = d ( 2)(m t);d = , where @xmath57 is the dilution factor .
furthermore , i_0= 2(0)=2(0).[i0 ] the @xmath58 term can be probed by integrating over half the range of the variable @xmath45 which can be taken for instance to be @xmath59 for @xmath5 decay and @xmath60 for @xmath61 decay .
= & & + [ + ( m t ) + 2 ( 2)(m t ) + - 2 ( 2)(m t ) ] , + = & & + [ + ( m t ) - 2 ( 2)(m t ) + - 2 ( 2)(m t ) ] , where @xmath62 , @xmath63 , @xmath64 and @xmath65 , are the integrated @xmath54 , @xmath66 , @xmath55 and @xmath67 functions integrated over the range @xmath59 .
one can measure @xmath58 by fitting to the time distribution of @xmath68 .
the fits to the data yield @xcite & = & 0.760.18(stat)0.07(syst ) , + & = & 0.100.24 ( stat ) 0.06 ( syst ) , + & = & 0.380.24 ( stat ) 0.05 ( syst ) . using the theoretical calculation of @xmath69 @xcite @xmath19
is preferred to be positive at the 94% confidence level as the theoretical parameter @xmath70 is positive .
it is also interesting to note that @xmath71 is significantly different from zero .
this implies that there is a sizable resonant contribution to the decay from a unknown @xmath72 state with large width .
this can have important implications for the interpretation of the new @xmath73 states discovered recently @xcite .
the decay @xmath74 has been recently proposed to measure @xmath18 and @xmath19 @xcite .
the idea in this case is also to use intermediate resonances to probe @xmath19 . consider the decays of the @xmath5 meson to @xmath75 .
the dominant decay is through the @xmath76 quark level process .
the decay @xmath77 .,scaledwidth=40.0% ] next consider the multi body @xmath57 meson decay . as an example use @xmath78 .
we can write , amp(|d^0 ) & = & f(m_+^2,m_-^2 ) , + amp(d^0 ) & = & f(m_-^2,m_+^2 ) , + m_+,-^2=(p_k_s+p_^)^2.[ampd0 ] the amplitude for the @xmath5 decay at time @xmath79 is then given by , m_|b^0 ( ) & = & [ |b^0 d^0(k_s ^+ ^- ) h^0 ] c - + & & i _
h^0 ( -1)^l [ b^0 |d^0(k_s ^+ ^- ) h^0 ] s , + where @xmath80 and @xmath81 and [ eq : m_b0bar_pre ] m_|b^0 ( ) & = & f(m_-^2,m_+^2 ) c - + & & i _ h^0 ( -1)^l f(m_+^2,m_-^2 ) s. then , using eq . [ ampd0 ] , [ eq : m_b0bar ] m_|b^0 ( ) & = & f(m_-^2,m_+^2 ) c - + & & i e^-i2_1 _ h^0 ( -1)^l f(m_+^2,m_-^2 ) s , + [ eq : m_b0 ] m_b^0 ( ) & = & f(m_+^2,m_-^2 ) c - + & & i e^+i2_1 _ h^0 ( -1)^l f(m_-^2,m_+^2 ) s. one then needs to model @xmath82 to extract the phase @xmath83 .
this is done by assuming breit - wigner form for the various contributing resonances .
a non resonant component is also included .
the experimental measurements of this process is discussed in other talks in the workshop@xcite .
the decay @xmath85 .,scaledwidth=40.0% ] the amplitude for @xmath86 can be written as , a^d & = & ( t + e + p_c ) v_cb^ * v_cd + p_u v_ub^ * v_ud + + & & ( p_t + p_ew^c ) v_tb^ * v_td , + & = & ( t + e + p_c - p_t - p_ew^c ) v_cb^ * v_cd + & & + ( p_u - p_t - p_ew^c ) v_ub^ * v_ud , + & & e^i ^ct + e^i e^i ^ut .
[ ampddbar ] the time - dependent measurement of this decay allows one to obtain the three observables , b & & ( |a|^2 + & + & 2 , + a_dir & & ( |a|^2 - |a|^2 ) = - 2 , + a_indir & & i m ( e^-2i a^ * a ) + & - & ^2 2 - 2 ( 2 + ) + & - & ^2 ( 2 + 2 ) . the three independent observables depend on five theoretical parameters : @xmath87 , @xmath88 , @xmath89 , @xmath2 , @xmath90 .
if @xmath87 can be neglected then one can obtain @xmath18 .
in general , however , one can not obtain cp phase information from these measurements- this is the well known penguin pollution problem .
hence theoretical input is necessary to get the ckm phase information .
one can partially solve for the theory amplitudes as , ^2 & = & a_r ( 2 + 2 ) - a_indir ( 2 + 2 ) - b 2- 1 , +
a_r^2 & = & b^2 - a_dir^2 - a_indir^2 . one can then obtain @xmath88 from a partner process and use it as a theory input .
we can then obtain @xmath90 if given @xmath2 and vice versa .
the partner process can be chosen to be @xmath91 @xcite or @xmath92 @xcite .
the measurement of both @xmath18 and @xmath19 is crucial to test the standard model predictions of cp violation .
this is all the more important as there are several decays where there are hints of beyond standard model physics@xcite , making the precise measurements of the ckm angles a top priority . in this talk
we have described several decays involving b decays to charm that probe both @xmath18 and @xmath19 .
t. e. browder , a. datta , p. j. odonnell and s. pakvasa , phys .
d * 61 * , 054009 ( 2000 ) [ arxiv : hep - ph/9905425 ] ; for earlier papers see : j .
charles , a. le yaouanc , l. oliver , o. pene and j. c. raynal , phys .
b * 425 * , 375 ( 1998 ) [ erratum - ibid .
b * 433 * , 441 ( 1998 ) ] [ arxiv : hep - ph/9801363 ] ; p. colangelo , f. de fazio , g. nardulli , n. paver and riazuddin , phys .
d * 60 * , 033002 ( 1999 ) [ arxiv : hep - ph/9901264 ] .
a. datta , int .
j. mod .
a * 19 * , 5501 ( 2004 ) [ arxiv : hep - ph/0407330 ] ; a. datta and p. j. odonnell , phys .
b * 572 * , 164 ( 2003 ) [ arxiv : hep - ph/0307106 ] ; a. datta and p. j. odonnell , phys .
b * 567 * , 273 ( 2003 ) [ arxiv : hep - ph/0306097 ] .
a. bondar , t. gershon and p. krokovny , phys .
b * 624 * , 1 ( 2005 ) [ arxiv : hep - ph/0503174 ] .
see for instance , k. a. george , arxiv : hep - ex/0701044 .
j. albert , a. datta and d. london , phys .
b * 605 * , 335 ( 2005 ) [ arxiv : hep - ph/0410015 ] ; a. datta and d. london , phys .
b * 584 * , 81 ( 2004 ) [ arxiv : hep - ph/0310252 ] .
r. fleischer , eur .
j. c * 10 * , 299 ( 1999 ) [ arxiv : hep - ph/9903455 ] .
see for instance , s. baek , a. datta , p. hamel , o. f. hernandez and d. london , phys .
d * 72 * , 094008 ( 2005 ) [ arxiv : hep - ph/0508149 ] ; s. baek , p. hamel , d. london , a. datta and d. a. suprun , phys .
d * 71 * , 057502 ( 2005 ) [ arxiv : hep - ph/0412086 ] ; a. datta , m. imbeault , d. london , v. page , n. sinha and r. sinha , phys .
d * 71 * , 096002 ( 2005 ) [ arxiv : hep - ph/0406192 ] ; a. datta and d. london , phys .
b * 595 * , 453 ( 2004 ) [ arxiv : hep - ph/0404130 ] ; a. datta , phys .
d * 66 * , 071702 ( 2002 ) [ arxiv : hep - ph/0208016 ] ; a. datta , x. g. he and s. pakvasa , phys .
b * 419 * , 369 ( 1998 ) [ arxiv : hep - ph/9707259 ] ; t. e. browder , a. datta , x. g. he and s. pakvasa , phys .
d * 57 * , 6829 ( 1998 ) [ arxiv : hep - ph/9705320 ] . | the measurement of @xmath0 from @xmath1 does not resolve the discrete ambiguities in the angle @xmath2 .
a measurement of @xmath3 is therefore desirable .
this talk is about measuring the ckm angle @xmath4 from various @xmath5 decays to charm final states which also allow us a measurement of @xmath3 . |
the study of the properties of nuclei far from the line of @xmath4 stability is currently one of the most active and challenging topics in nuclear structure .
exotic ratios of z / n for a given mass number together with the presence of weakly bound nucleons lead to the phenomena which are unexpected from our common sense about stable nuclei ; among others , one finds the change of the shell structure and magic numbers . observed properties of low - lying states in @xmath5be@xmath6 and @xmath7be@xmath8 have contributed to the finding that n=8 is no longer a magic number in such unstable nuclei . though there have been already many elaborate attempts to describe the structure of be isotopes ,
in the present paper we give an interpretation of available spectroscopic data on @xmath2be and @xmath3be keeping a model as simple as possible , while taking duly into account the essential feature of these nuclei ; the presence of weakly - bound neutron(s ) in finite deformed potentials . a simplest version of bohr - mottelson model ( see chaps . 4 and 5 of ref .
@xcite ) is applied to the nuclei @xmath3be and @xmath2be , using weakly - bound one - particle wave - functions estimated properly with deformed woods - saxon potentials instead of harmonic - oscillator potentials .
the present work is prompted by the recent lifetime measurements of the 0@xmath9 state in @xmath2be @xcite .
the deformation of be isotopes has been studied using various theoretical models ; for example , @xmath3be is studied using models with deformed woods - saxon potentials @xcite , @xmath3be based on the nilsson - strutinsky method @xcite , and neutron - rich be isotopes using the deformed hartree - fock calculation with skyrme interactions @xcite .
the models used in refs .
@xcite , in which the idea of the particle - rotor model is used , may be closest to our present model .
however , since in very light nuclei such as be isotopes the nuclear shape may be considerably changed just by adding one neutron , we would avoid to use the observed properties of the core nucleus @xmath10be , in contrast to the work of refs . @xcite .
since only low - energy states are discussed in the present work , we assume that four protons and six neutrons are treated as being inactive occupying the lowest possible nilsson levels for the same deformation as that for the extra neutron(s ) .
thus , the inactive core is certainly different from the actual @xmath10be nucleus .
moreover , the rotational perturbation of the intrinsic nuclear structure is neglected since only those states close to the band head are studied in the present work .
thus , we try to see how the observed properties of low - lying levels of @xmath3be and @xmath2be can be understood in terms of weakly - bound neutron(s ) in deformed potentials .
deformed nuclear halos are studied in ref .
@xcite using a finite square - well potential without spin - orbit term , while in ref .
@xcite they are investigated with more realistic potentials : deformed woods - saxon potentials with a spin - orbit term . in sec .
ii some aspects of the model and formulas are summarized . in sec.iii numerical results
are presented and discussed in comparison with experimental data , while in sec .
iv conclusions are given .
it is known @xcite that in medium - heavy deformed nuclei the analysis of low - energy spectroscopic properties in terms of one - particle motions in a deformed potential works impressively well , especially if the pair - correlation in the form of bcs quasiparticles is included .
this is because the major part of the long - range two - body interaction is already taken into account in the deformed mean - field . in very light nuclei such as
be isotopes the many - body pair correlation that originates from a short - range interaction can be negligible . therefore ,
in the present work we try to describe the spectroscopic properties of low - lying states of @xmath3be and @xmath2be in terms of two nilsson levels occupied by the seventh neutron and the seventh and eighth neutrons , respectively , in deformed woods - saxon potentials .
the pairing interaction between two weakly - bound neutrons in @xmath2be will be duly taken into account .
this simple description may be justified , since there are no nearby nilsson levels in the region of deformation , where the [ 220,1/2 ] and [ 101,1/2 ] levels are almost degenerate . one - particle levels with quantum numbers k@xmath11 in a @xmath12-deformed potential ( nilsson levels ) are denoted by the asymptotic quantum numbers , [ @xmath13 , where @xmath14 and @xmath15 .
the asymptotic quantum numbers express , in a good approximation , the wave functions of one - particle levels well bound in sufficiently deformed potentials .
when a one - particle level becomes weakly bound , the wave function can be very different from that indicated by the asymptotic quantum numbers , even when the one - particle potential is well deformed @xcite .
nevertheless , just for convenience sake , in the present work we use the asymptotic quantum numbers to denote respective nilsson levels , also in the case that those levels become weakly bound . in the leading order
the rotational energy spectrum is written as @xmath16 where @xmath17 expresses the moment of inertia while the decoupling parameter is denoted by @xmath18 . the expression ( [ eq : erot ] ) , in which the rotational perturbation of the intrinsic structure is neglected , is expected to work well for low - lying rotational states close to the band head .
the formulas for b(e@xmath19 ) to be used in the present work are ( see eqs .
( 4 - 91 ) and ( 4 - 92 ) of ref .
@xcite ) @xmath20 and @xmath21 1 & \qquad \mbox{for } & \qquad k_2 = 0 \end{array } \right .
\label{eq : belam0}\end{aligned}\ ] ] where intrinsic matrix - elements are expressed by @xmath22 . from two different pairs of one - particle orbits ( @xmath23 ) and ( @xmath24 ) in deformed potentials , where @xmath25 expresses the time - reversed orbit of @xmath26 , one can form two orthogonal two - particle configurations @xmath27 this mixture of the two @xmath28 configurations can be interpreted as the result of pairing interaction .
the quadrupole matrix element connecting these two states in the intrinsic system is written as @xmath29 similarly , the monopole matrix element is @xmath30 the unique feature of these transition matrix - elements is illuminated on p.552 - 553 of ref .
@xcite , in connection with pair correlation and @xmath4 vibration . in order to obtain the wave functions of one - particle levels in deformed woods - saxon potentials ,
the coupled - channel equations for a given one - particle level are solved @xcite in coordinate space with correct asymptotic conditions for @xmath31 , without confining the system in a finite box .
in @xmath3be known as a one - neutron halo nucleus the presence of an extremely strong e1 transition , @xmath32 = @xmath33 e@xmath34fm@xmath34 , is well known for years @xcite .
there are only two known bound states in @xmath3be : the 1/2@xmath36 ground state and the 1/2@xmath37 state at ex=0.32 mev . due to the very small neutron separation energy @xmath38=0.504 mev
, there is no hope to observe enhanced e2 transitions to establish possible rotational spectra based on the ground state .
however , at least three resonant levels , 5/2@xmath39 at ex = 1.78 mev , 3/2@xmath40 at ex = 2.69 mev and 3/2@xmath39 at ex = 3.41 mev , have been established by recent experiments @xcite .
if we regard the two one - particle resonant levels , 5/2@xmath39 at ex = 1.78 mev and 3/2@xmath39 at ex = 3.41 mev , as members of the rotational band based on the ground state with @xmath41 = 1/2@xmath36 , we obtain the decoupling parameter @xmath18=1.82 and the rotational constant @xmath42=0.403 mev .
then , using the expression in eq .
( [ eq : erot ] ) , the member of the ground - state rotational band in @xmath3be next lowest to 3/2@xmath36 is the 9/2@xmath39 state , which is expected around ex=6.7 mev in the absence of rotational perturbation .
the centrifugal barrier for the 9/2@xmath39 level is quite high and , thus , the possible resonance will be relatively sharp in spite of its higher energy .
assuming that the moment of inertia @xmath17 is proportional to @xmath43 , @xmath42=0.349 mev is obtained for @xmath2be .
then , we estimate the excitation energy of the 2@xmath44 state in @xmath2be to be @xmath45 which indeed agrees with the observed value 2.11 mev .
this agreement may indicate that the two nuclei , @xmath2be and @xmath3be , have similar deformations . in @xmath3be
we assume that the 1/2@xmath39 and 1/2@xmath40 levels are expressed by the seventh neutron occupying one - neutron levels , [ 220,1/2 ] and [ 101,1/2 ] , respectively , for a given strong prolate - deformation . using the standard spin - orbit strength @xcite , radius parameter @xmath46=1.25 fm , and a diffuseness @xmath18=0.65 fm ( potential [ a ] ) or @xmath18=1.00 fm ( potential [ b ] ) , the deformation @xmath4 and the depth of the woods - saxon potential @xmath47
are determined so as to produce @xmath48([220,1/2])[email protected] mev and @xmath48([101,1/2])[email protected] mev .
we obtain @xmath50 } \\
\beta = 0.82 , \qquad v_{ws } = -33.5 \ : \mbox{mev } \qquad & \mbox{for potential [ b ] } \end{aligned}\ ] ] the @xmath4 value determined for @xmath3be in the literature depends on the used model and parameterization , and varies from 0.7 to 1.1 ; for example , see ref .
the diffuseness @xmath18=0.65 fm is a standard value for medium - heavy stable nuclei , while @xmath18=1.00 fm is an attempt to simulate the potential without a flat bottom that may be commonly obtained in hartree - fock calculations for very light nuclei . in the present work
the deformed woods - saxon potential is needed only to obtain the wave function of the seventh neutron , though the total neutron potential in the presence of a halo neutron may not be properly described by a woods - saxon potential . in table
[ tab : table1 ] we show the probabilities of @xmath51 components in the wave functions of the [ 220,1/2 ] and [ 101,1/2 ] levels estimated for the potentials [ a ] and [ b ] .
it is noted that the radial wave - functions of @xmath51 components of those deformed weakly - bound levels can be considerably different from those of @xmath52 eigenfunctions of spherical woods - saxon potentials @xcite .
therefore , great care has to be taken , when the spectroscopic factors are extracted in the analysis of the data such as the one made in refs . @xcite . in this sense
, a caution is needed in the comparison of the @xmath53 probability , 0.62 , in the [ 220,1/2 ] level calculated for the potential [ b ] with the spectroscopic factor obtained in ref .
@xcite , [email protected] . using the [ 220,1/2 ] neutron wave - function for the potential [ b ] with @xmath55=z / a and @xmath56=0.9@xmath57
, we obtain @xmath491.72 @xmath58 for the magnetic moment of the ground state of @xmath3be , which is in good agreement with the observed value , @xmath59(@xmath3be ) = @xmath491.6816(8 ) @xmath58 @xcite .
the calculated value is not sensitive either to the used @xmath55 value of the even - even core or the parameters of the woods - saxon potential .
this insensitivity comes from the fact that the magnetic moment of an @xmath53 neutron is equal to 0.5@xmath56 , while in the asymptotic limit the magnetic moment of the i=1/2 state coming from the [ 220,1/2 ] neutron coupled to the k=0 core is also equal to 0.5@xmath56 independent of the core @xmath55-value used . in other words , in the present i=1/2 state the value of the measured magnetic moment can not be used to find whether or not the nucleus is deformed .
the spectroscopic properties of four known bound levels in @xmath2be , @xmath60 , @xmath61 at ex=2.11 mev , @xmath62 at ex=2.25 mev and @xmath63 at ex=2.70 mev , have been experimentally studied for the last few years @xcite . in particular , the recent measurement of the lifetime of the @xmath62 level @xcite has pinned down the absolute magnitudes of @xmath64 = ( [email protected] ) e@xmath65 fm@xmath65 and @xmath67 = ( [email protected] ) e fm@xmath65 , which are very valuable informations .
the relatively large values of @xmath64 and @xmath67 indicate that the intrinsic configurations of the @xmath60 and @xmath62 states are expressed by linear combinations of two two - neutron configurations , ( @xmath68 ) and ( @xmath69 ) , of which the intrinsic quadrupole moments are quite different .
the analysis of inelastic proton scattering exciting the @xmath61 state in inverse kinematics @xcite led to the deformation length , ( [email protected] ) fm , which suggested strong quadrupole deformation , @xmath70
. observed two 0@xmath39 levels in @xmath2be are assumed to be expressed by linear combinations ( with equal amplitudes ) of two two - neutron configurations , [ 220,1/2]@xmath71}$ ] and [ 101,1/2]@xmath72}$ ] , for a given prolate deformation .
the relation @xmath73 in the expression of eq .
( [ eq : ab ] ) is expected since the [ 202,1/2 ] and [ 101,1/2 ] levels in @xmath3be lie only 300 kev apart , following the interpretation in the present model .
the deformation @xmath4 and the depth of the woods - saxon potential @xmath47 are determined so as to produce both [ 220,1/2 ] and [ 101,1/2 ] levels at @xmath491.1 mev , which in the absence of the interaction between the two configurations is about a half of the binding energy of the two neutrons .
then , we obtain @xmath74 } \\ \beta = 0.72 , \qquad v_{ws } = -35.2 \ : \mbox{mev } \qquad & \mbox{for potential [ b ] } \end{aligned}\ ] ] the intrinsic configuration assumed here for the ground state of @xmath2be is consistent with the fact that all four states , 1/2@xmath36 , 1/2@xmath37 , 5/2@xmath36 and 3/2@xmath37 , of @xmath3be are populated in one - neutron removal reactions from @xmath2be with comparative magnitudes of spectroscopic factors @xcite
. in medium - heavy stable nuclei low - energy e1 transitions are so much hindered in both spherical and deformed nuclei that one could hardly expect to obtain a nuclear - structure information from observed b(e1 ) values @xcite .
the reason for the hindrance is : ( a ) no appreciable amount of low - energy e1 strength due to the nuclear shell - structure ;
( b ) the high - lying isovector giant dipole resonance absorbs the major part of possible e1 strength ; ( c ) isoscalar dipole mode corresponds to the center of mass motion and thus does not contribute to e1 excitations .
in contrast , in nuclei with weakly bound nucleons the origin of the hindrance ( a ) may be removed due to the change of the shell - structure or change in the one - particle wave functions , while the hindrance ( b ) may be drastically reduced due to the very weak coupling of the nucleons to the well - bound core . in
well - bound and well - deformed nuclei the e1 transition between the [ 220,1/2 ] and [ 101,1/2 ] levels , where @xmath75=2 and @xmath76=1 , is asymptotically forbidden .
therefore , if the nucleus is well deformed , the observed strong e1 transition must come from the properties of the levels related to being weakly bound .
it is pointed out that the @xmath53 component becomes dominant in all @xmath77 nilsson levels as the binding energy of the orbits approaches zero , though it depends on respective nilsson levels at which binding energy the dominance occurs @xcite .
since the major component of the [ 101,1/2 ] level is @xmath78 , the @xmath53 dominance in the weakly - bound [ 220,1/2 ] level can produce a pretty strong e1 matrix - element . in table
[ tab : table2 ] the calculated result of e1 properties is shown , where the neutron e1 effective charge , @xmath79 = @xmath80 = ( 0.364)e , is used neglecting the core polarization effect ( namely , the effect of shifting low - energy e1 strength to the high - lying isovector giant dipole resonance ) since halo neutrons may hardly polarize the well - bound core .
the formula used to estimate the b(e1 ) value is @xmath81 | ry_{10 } | [ 220,1/2 ] \rangle \nonumber \\ & & \mbox { } + ( -1)^{1/2 + 1/2 } c(\frac{1}{2},1 , \frac{1}{2 } ; \frac{-1}{2},1 , \frac{1}{2 } ) \langle [ 101,1/2 ] | ry_{11 } | [ \widetilde{220,1/2 } ] \rangle \}^2 \end{aligned}\ ] ] where @xmath82 $ ] expresses the time - reversed intrinsic configuration of @xmath83 $ ] . in the present case
the contribution coming from the @xmath84 term to the e1 matrix element ( so - called signature - dependent term ) , which originates from the @xmath85 symmetry of the nuclear shape , is several times larger than that from the @xmath86 term and contributes coherently . in both @xmath86 and @xmath84 matrix elements
a smaller contribution from @xmath87 contributes destructively to the major one from @xmath88 .
the destructive structure is a remaining trace of the asymptotically forbidden e1 matrix - element . observing that the b(e1 ) value obtained by using the potential [ b ] is in good agreement with the observed value , ( [email protected] ) e@xmath34 fm@xmath65 , in the following we show numerical results obtained with the potential [ b ] . in the e1 transition
, @xmath89 , of @xmath2be the @xmath86 term contributes if the @xmath90 state has @xmath91 , while the @xmath84 term does if @xmath92 . in both @xmath91 and @xmath92
cases the band head will be @xmath90 . if @xmath91 , the rotational band has members with @xmath93 , while the rotational band with @xmath92 has @xmath94 that means , if one observes the 2@xmath37 rotational member among resonant levels , @xmath95 is assigned to the @xmath96 level at ex=2.70 mev . since , to our knowledge , no resonant level that may be a candidate for the rotational member
is yet found , in table [ tab : table3 ] we show the calculated results of both @xmath97 and @xmath98 cases .
the intrinsic wave functions used are @xmath99[\widetilde{220,1/2 } ] + [ 101,1/2][\widetilde{101,1/2 } ] ) \nonumber \\ \phi ( k^{\pi } = 0 ^ - ) & = & \frac{1}{\sqrt{2 } } ( [ 101,1/2][\widetilde{220,1/2 } ] - [ \widetilde{101,1/2}][220,1/2 ] ) \nonumber \\ \phi ( k^{\pi } = 1 ^ - ) & = & [ 101,1/2][220,1/2 ] \label{eq : wf01}\end{aligned}\ ] ] then , the formulas used to estimate those b(e1 ) values are @xmath100| ry_{10 } | [ 220,1/2 ] \rangle |^2 ( e_{eff}^{n } ( e1))^2 \nonumber \\ \\ b(e1;i^{\pi } = k^{\pi } = 0_1^+ \rightarrow i^{\pi } = k^{\pi } = 1 ^ - ) = 4 | \langle [ 220,1/2]| ry_{11 } | [ \widetilde{101,1/2 } ] \rangle |^2 ( e_{eff}^{n } ( e1 ) ) ^2 \nonumber \\\end{aligned}\ ] ] in table [ tab : table3 ] @xmath101e = ( 4/12)e = ( 0.333)e is employed . since the matrix - element of @xmath102 is larger than that of @xmath103 , the calculated b(e1;@xmath104 ) value is several times larger than the b(e1;@xmath105 ) value . since @xmath106 for
e1 transitions for neutrons with the binding energy of @xmath491.1 mev may be somewhat smaller than ( z / a)e due to non - negligible core - polarization effect , the assignment of @xmath95 to the @xmath107 state at ex=2.70 mev is suggested from the comparison of the calculated and observed b(e1 ) values .
it is noted that in the present simple model , where @xmath108 is assumed in the expression ( [ eq : ab ] ) , both the b(e1 ; @xmath109 ) and b(e1 ; @xmath110 ) values in @xmath2be vanish .
thus , it is very interesting to measure the upper limit of the e1 transition , 1@xmath37 at 2.70 mev @xmath111 0@xmath112 at 2.25 mev , though the small transition - energy makes the measurement very difficult .
now , we consider the e2 transition , 0@xmath9 @xmath113 2@xmath44 , in @xmath2be .
the 2@xmath114 state at ex = 2.11 mev is interpreted as a rotational member of the ground 0@xmath44 state . using the intrinsic configuration @xmath115[\widetilde{220,1/2 } ]
+ [ 101,1/2][\widetilde{101,1/2 } ] ) \label{eq : wf02}\end{aligned}\ ] ] and @xmath116 in eq .
( [ eq : wf01 ] ) , we obtain @xmath117 | r^2y_{20 } | [ 220,1/2 ] \rangle - \langle [ 101,1/2 ] | r^2y_{20 } | [ 101,1/2 ] \rangle \}^2 \label{eq : e200}\end{aligned}\ ] ] using the potential [ b ] for @xmath2be we obtain @xmath118 | r^2y_{20 } | [ 220,1/2 ] \rangle = 11.56 \quad \mbox{fm}^2 \nonumber \\ \langle [ 101,1/2 ] | r^2y_{20 } | [ 101,1/2 ] \rangle = -0.32 \quad \mbox{fm}^2 \label{eq : r2y2}\end{aligned}\ ] ] and , then , @xmath119 where @xmath120 = ( 0.2 @xmath121 0.3)e is used in the last line .
these values of @xmath120 are obtained from the analysis of measured quadrupole moments of @xmath122b@xmath123 and @xmath124b@xmath125 using the shell model with the @xmath126 space @xcite .
both the large neutron excess and the weak binding of neutrons make @xmath120 smaller .
the estimated value ( [ eq : be2 ] ) is in agreement with the measured value , b(e2 ; @xmath127 ) = ( 7.0 @xmath54 0.6 ) e@xmath65@xmath128 .
it is noted that for the present large prolate deformation the matrix element @xmath129 | r^2 y_{20 } | [ 101,1/2 ] \rangle$ ] is positive if the [ 101,1/2 ] level is deeply bound , in contrast to the negative value in ( [ eq : r2y2 ] ) . on the other hand , for a more moderate prolate deformation
the matrix element is negative irrespective of one - particle energies .
the dependence of the expectation values of @xmath130 in the [ 101,1/2 ] and [ 220,1/2 ] states on one - particle energies is exhibited in fig . 1 for two deformation values .
it is seen that in the region of @xmath131 mev the indication of the divergence , @xmath129 | r^{2 } y_{20 } | [ 101,1/2 ] \rangle$ ] @xmath113 @xmath132 for @xmath133 , which comes from the behavior of @xmath134 one - particle wave - functions for @xmath135 @xcite , has not yet appeared .
thus , in the present case the calculated value of the expression ( [ eq : e200 ] ) is not sensitive to a small variation of @xmath136 .
the very small value of @xmath137 \mid r^2 y_{20 } \mid [ 101,1/2 ] \rangle \mid$ ] for @xmath131 mev comes from the dominance of p@xmath138 component in the [ 101,1/2 ] wave function , because the density distribution of p@xmath138 orbit is spherically symmetric . on the other hand , for all values of @xmath139 the major contribution to @xmath129 | r^{2 } y_{20 } | [ 101,1/2 ] \rangle$ ] comes from the @xmath140 matrix element between the components @xmath78 and @xmath141 in the [ 101,1/2 ] wave function . note that the ratio of the probability of @xmath141 to that of @xmath78 in the [ 101,1/2 ] wave function has no strong variation for @xmath135 . in order to see the effect of weakly - bound neutrons
, we estimate the above b(e2 ) value using the wave functions in deformed harmonic - oscillator potentials ( see , for example , ref .
@xcite ) . noting that @xmath118 | 2z^2 - x^2 - y^2 | [ 220,1/2 ] \rangle = 5 c_{z}^2 - c_{\perp}^2 \\ \langle [ 101,1/2 ] | 2z^2 - x^2 - y^2 | [ 101,1/2 ] \rangle = c_{z}^2 - 2 c_{\perp}^2 \label{eq : czp}\end{aligned}\ ] ] in the unit of @xmath142 , where @xmath143 and @xmath144 expresses the @xmath12 deformation parameter of the oscillator potential , we obtain @xmath118 | r^2y_{20 } | [ 220,1/2 ] \rangle & - & \langle [ 101,1/2 ] | r^2y_{20 } | [ 101,1/2 ] \rangle \nonumber \\ & = & \sqrt{\frac{5}{16 \pi } } \ , ( 4 c_{z}^2 + c_{\perp}^2 ) \qquad \mbox{in unit of } \quad \hbar / m \omega_{0}(\epsilon ) \nonumber \\ & = & 2.62 \quad \mbox{in unit of } \quad \hbar / m \omega_{0}(\epsilon ) \nonumber \\ & = & 5.66 \quad \mbox{fm}^2 \qquad \mbox{for a=12 } \label{eq : tr2y2}\end{aligned}\ ] ] where @xmath144=0.7 is used and the volume conservation condition @xmath145 is taken into account with @xmath146 = 41 a@xmath147 mev . using eqs .
( [ eq : r2y2 ] ) and ( [ eq : tr2y2 ] ) it is seen that the weakly - bound character of two neutrons in @xmath2be makes the b(e2;@xmath148 ) value larger by ( 11.88/5.66)@xmath65 = 4.41 .
it is noted that the matrix element @xmath129|r^2 y_{20}|[101,1/2 ] \rangle$ ] in the oscillator potential , which is given in ( [ eq : czp ] ) , changes the sign at @xmath144=0.6 and is positive for @xmath144=0.7 , in agreement with the sign change of the @xmath140 matrix element of the deeply - bound [ 101,1/2 ] level shown in fig . 1 . for the e0 transition , 0@xmath9 @xmath113 0@xmath44 , in @xmath2be we obtain @xmath149| r^2 |[220,1/2 ] \rangle = 34.4 \
: \mbox{fm}^2 \nonumber \\ \langle [ 101 , 1/2]| r^2 |[101,1/2 ] \rangle = 23.0 \ : \mbox{fm}^2\end{aligned}\ ] ] by using the wave functions calculated with the potential [ b ] for @xmath2be .
then , using the intrinsic wave - functions , ( [ eq : wf01 ] ) and ( [ eq : wf02 ] ) , we obtain @xmath150 which is the neutron matrix - element .
the neutron effective charge for e0 transitions , @xmath151 , may sensitively depend on both neutron excess and weak binding .
the contribution to @xmath151 coming from the part of the one - body operator obtained by subtracting the center of mass motion is @xmath152 , which is equal to ( 0.028)e for @xmath2be
. however , this may hardly give a reliable estimate of the actual @xmath151 value .
if the charge radius of @xmath2be relative to that of either @xmath3be or @xmath10be is ever measured , it will help to obtain an estimate of the @xmath151 value to be used in the present case @xcite .
since no such measurement is presently available , writing the measured value as @xmath153 we obtain @xmath154
we have presented the interpretation of available spectroscopic data on @xmath2be and @xmath3be , using a model which is simple and contains the essential feature of the presence of weakly - bound neutron(s ) in deformed potentials .
calculated results are in good agreement with available data .
we have intentionally avoided to make a quantitative comparison of spectroscopic factors with `` measured '' ones , partly because the presently available factors obtained from the analysis of data may , in our opinion , contain an ambiguity in the one - particle radial wave - functions used and partly because the inactive core nucleus @xmath10be in our present model is different from the observed nucleus @xmath10be .
the features of weakly - bound neutrons appear especially in : ( a ) strong e1 transitions in both @xmath3be and @xmath2be due to the increased and spatially - extended @xmath53 component in the [ 220,1/2 ] wave function ; ( b ) the large b(e2;@xmath155 ) value , due to both the larger value of @xmath156 | r^2 y_{20 } | [ 220,1/2 ] \rangle$ ] and the negative value of @xmath129 | r^2 y_{20 } | [ 101,1/2 ] \rangle$ ] ; ( c ) the large @xmath156 | r^2 | [ 220,1/2 ] \rangle$ ] matrix element that contributes to the e0 transition , @xmath157 . in medium - heavy mass region one
hardly finds a deformed nucleus , in which an @xmath158=1/2@xmath39 level lies close to an @xmath158=1/2@xmath37 level around the ground state .
therefore , the nucleus @xmath3be is the rare case , in which @xmath158=1/2@xmath36 and 1/2@xmath37 levels are almost degenerate and the importance of the presence of the signature - dependent term in b(e1 ) values is manifested .
99 a. bohr and b. r. mottelson , _ nuclear structure _
( benjamin , reading , ma , 1975 ) , vol.ii . s. shimoura et al . , submitted to phys .
lett . * b*. h. esbensen et al .
c * 51 * , 1274 ( 1995 ) .
a. muta and t. otsuka , prog .
, * 142 * , 355 ( 2001 ) .
i.ragnarsson , s. berg , h .- b .
hkansson and r. k. sheline , nucl . phys . *
a361 * , 1 ( 1981 ) .
x. li and p .- h.heenen , phys .
c * 54 * , 1617 ( 1996 ) .
t. misu , w. nazarewicz and s. berg , nucl . phys . *
a614 * , 44 ( 1997 ) .
i. hamamoto , phys .
c * 69 * , 041306 ( 2004 ) .
d. j. millener , j. w. olness , e. k. warburton and s. s. hanna , phys . rev .
c * 28 * , 497 ( 1983 ) . n. fukuda et al . , phys .
c * 70 * , 054606 ( 2004 ) .
s. d. pain et al .
* 96 * , 032502 ( 2006 ) .
a. bohr and b. r. mottelson , _ nuclear structure _
( benjamin , reading , ma , 1969 ) , vol.i . i. hamamoto , phys . rev .
c * 72 * , 024301 ( 2005 ) .
et al . , phys .
lett . * 85 * , 266 ( 2000 ) .
w. geithner et al .
lett . * 83 * , 3792 ( 1999 ) .
h. iwasaki et al . , phys .
lett . * b481 * , 7 ( 2000 ) .
h. iwasaki et al . , phys . lett . * b491 * , 8 ( 2000 ) .
s. shimoura et al . , phys . lett . * b560 * , 31 ( 2003 ) .
for example , see ; g. b. hagemann , i. hamamoto and w. satula , phys .
c * 47 * , 2008 ( 1993 ) .
h. sagawa and k. asahi , phys .
c * 63 * , 064310 ( 2001 ) . k. riisager , a. s. jensen and p. mller , nucl . phys . * a548 * , 393 ( 1992 ) .
j. p. boisson and r. piepenbring , nucl . phys . * a168 * , 385 ( 1971 )
. b. r. mottelson , private communication .
.[tab : table1 ] calculated probabilities of @xmath51 components in the [ 220,1/2 ] and [ 101,1/2 ] wave functions of @xmath3be . calculated radial wave - functions of the @xmath51 components are in general different from those of the ( @xmath159 ) wave functions which are the eigenfunctions of spherical woods - saxon potentials .
two parameters of a given potential , deformation @xmath4 and the depth of woods - saxon potentials , are adjusted so that the eigen energies of the [ 220,1/2 ] and [ 101,1/2 ] orbits are @xmath490.5 and @xmath490.2 mev , respectively .
see the text for details . [
cols="^,^,^,^,^,^,^,^,^,^,^ " , ] figure 1 : : : the expectation value of @xmath130 in the [ 101,1/2 ] and [ 220,1/2 ] states as a function of one - particle energy , @xmath136 .
for the [ 220,1/2 ] level the @xmath53 , @xmath160 , @xmath161 , @xmath162 and @xmath163 components are included in solving the coupled - channel equations for deformed one - particle levels , while for the [ 101,1/2 ] level the @xmath78 , @xmath141 , @xmath164 and @xmath165 components are taken into account . for comparison ,
two values of quadrupole deformation @xmath4 are considered .
used parameters are @xmath46 = 1.25 fm , @xmath18 = 1.0 fm and @xmath166 , while the depth of woods - saxon potentials is varied so that respective one - particle levels have @xmath136 as energy eigenvalues . | inspired by the recent measurement of the b(e2;@xmath0 ) and b(e0;@xmath1 ) values in @xmath2be , we give an interpretation of available spectroscopic data on both @xmath2be and @xmath3be , using a simple model which contains the essential feature of these two nuclei ; the presence of weakly - bound neutron(s ) in deformed potentials .
the agreement of the calculated results with observed data is surprisingly good , including well - known strong e1 transitions in both nuclei . |
the spatiotemporal patterns and transport of magnetic field in the photosphere are driven by the turbulent plasma flows in the solar convection zone .
solar magnetic fields are observed in complex and hierarchical structures covering a wide range of scales that emerge and vanish on the time - scales of turbulent convective patterns .
traditionally , they are classified by size and lifetime as patterns of granulation ( 1 mm , 0.2 h ) , mesogranulation ( 5 - 10 mm , 5 h ) and supergranulation ( 15 - 35 mm , 24 h ) .
recently , the idea is emerging that meso and supergranulation are signatures of a collective interaction of granular cells @xcite .
the continuous restructuring of surface magnetic fields by photospheric turbulent flows plays a key role in determining the topology and evolution of chromospheric and coronal magnetic fields , and may influence the triggering of eruptive solar events such as flares , coronal mass ejections or sudden disappearance of filaments @xcite .
for example , @xcite and @xcite used space and ground observations to show that large - scale horizontal photospheric flows below and around a filament influence the formation and evolution of filaments , leading to destabilization of the coronal magnetic field .
spectral line observations of the sun with high spatial resolution show small - scale bright points , with widths @xmath0 mm , located in the intergranular lanes .
bright points are associated with localized regions of strong magnetic field and correspond to magnetic flux tubes of kilogauss field strength that stand nearly vertically in the solar atmosphere @xcite .
some bright points exhibit the proper motions of convectively driven vortical flows created at the downdrafts where the plasma follows spiral paths and returns to the solar interior after cooling down @xcite .
recent balloon - borne sunrise observations show @xmath1 vortices mm@xmath2 min@xmath3 , with a mean lifetime of 7.9 min and a standard deviation of 3.2 min @xcite .
these magnetized vortex tubes have been predicted by the theory of magnetoconvection @xcite and observed in 3d radiative / compressible mhd simulations @xcite .
photospheric vortex structures with high concentration of magnetic field provide an important path for energy and momentum transfer from the convection zone into the chromosphere @xcite and may be relevant for chromospheric - coronal heating @xcite .
a number of observational and theoretical works have applied the lagrangian approach to trace the flow of bright points and the pattern of the magnetic network .
@xcite measured the horizontal flow field on the solar surface using the technique of local correlation tracking ( lct ) on a 28 min time sequence of white - light images of the solar optical universal polarimeter instrument onboard the spacelab2 .
they showed that insight into the relationship between flows and magnetic fields is acquired by calculating the flow of passive test particles , called `` corks '' , that are originally distributed uniformly in the flow field .
the cork paths , representing the lagrangian tracers of the pathlines of the flow , congregate at the same locations of the magnetic network .
a simplified kinematic model of convection at the solar surface was developed by @xcite to interpret observations and predict the evolution of magnetic fields .
they compared the cork patterns of the simulated flow with the cork patterns of the observed magnetic network , and discussed the extent to which magnetic flux tubes can be regarded as moving passively with the large - scale photospheric flow .
@xcite used the g - band images of the swedish vacuum solar telescope to show that bright points are arranged in linear structures , called `` filigrees '' , located in the lanes between neighboring granule cells and to measure the motion of bright points using an object tracking technique ; comparison with a 2d simulation of horizontal motions of magnetic flux elements in response to solar granulation flows shows that the observed velocities and spatial distribution of the bright points are consistent with passive advection of corks by the solar granulation flow .
@xcite reconstructed the 3d velocity field of a single supergranular cell from the spectrometer data of the interferometric bidimensional spectrometer , and used the tracking of corks to show that a divergence structure is created by a compact region of constant up - flow close to the cell centre , and isolated regions of strong convergent down - flow are nearby or cospatial with extended clusters of bright caii wing features forming the knots of the magnetic network .
@xcite applied the coherent structure tracking ( cst ) algorithm to track granules and analyze large - scale photospheric flows related to an extended filament in the active region ar 11106 using the @xmath4 image of the themis telescope and the euv imager ( aia ) of solar dynamic observatory .
they showed that diverging flows inside the supergranules may be similar in and out of the filament channel .
in addition , they identify converging flows , corresponding to accumulation of corks , around the footpoints or barbs located at the edges of the euv filament .
the frequent euv brightenings suggest the occurrence of reconnections of the magnetic field lines of the flux tube with the environment at the convergence points at the edges of the euv filament channel .
there is observational evidence of anomalous transport and turbulence in photospheric flows and magnetic fields . @xcite and @xcite applied the percolation theory to report sub - diffusion of magnetic elements in the photosphere .
@xcite used a 70 min sequence of g - band image of the swedish vacuum solar telescope on 5-october-1995 to show that the transport of bright points near the disk center is subdiffusive due to the trapping of walkers at the stagnation points in the fractal intercellular pattern ; the distribution of waiting times at the trap sites obeys a lvy ( power - law ) distribution .
@xcite applied subsonic filtering to the data of @xcite to report super - diffusion and show that the spatiotemporal scaling of the bright point dynamics indeed indicates the presence of turbulence in the intergranular lanes .
@xcite used a 32 min sequence of g - band and ca ii k - line intensity measurements of the dunn solar telescope on 28-may-2009 , at the disk center of size 32 mm @xmath5 32 mm , to investigate high - frequency fluctuations in both photosphere and chromosphere ; their noise - corrected g - band spectrum for @xmath6 mhz shows a power law with exponent @xmath7 , consistent with the presence of turbulent motions .
moreover , they showed that the g - band spectral power in the 25 - 100 mhz range is concentrated at the locations of magnetic bright points in the intergranular lanes and is highly intermittent in time , being characterized by a positive kurtosis , which implies that the fluctuations have non - gaussian probability distribution functions ( pdf ) with broad tails indicative of turbulence .
@xcite showed that the noise - filtered power spectrum of the hinode / sot data with a sequence of 80 min on 5-march-2007 , in the internetwork near the sun center of size 20 mm @xmath5 20 mm , presents a scaling range of @xmath8 mhz with characteristic power law exponents of -1.56 for g - band and of -4.00 for h - line , confirming the presence of turbulent fluctuations . by integrating the hinode / sot g - band spectra in the range 32 - 53 mhz , they identified the sites of increased g - band power with the location of g - band bright points .
@xcite and @xcite used the images of new solar telescope of big bear solar observatory to report super - diffusion and turbulent pair dispersion of bright points in active region plage , quiet - sun , and coronal hole .
the aim of this paper is to apply eulerian and lagrangian tools to detect coherent structures in images of photospheric turbulent flows .
first , we apply the eulerian approach to compute the q - criterion and distinguish regions of the plasma flow dominated by the deformation and vortical coherent structures @xcite .
next , we apply the tool of lagrangian coherent structures ( lcs ) to determine the repelling and attracting material lines that organize the transport of plasma flow .
the term `` lagrangian '' refers to flows defined by the fluid motion instead of an instantaneous snapshot ( eulerian ) ; the term `` coherent '' refers to the distinguished stability of these structures compared to other nearby material lines / surfaces .
there is a growing interest to adopt the lcs approach @xcite to improve the understanding of transport in complex flows such as planetary atmosphere @xcite , oceans @xcite , and cardiovascular system @xcite .
a number of papers have studied lcs in complex plasma flows for thermonuclear applications @xcite and astrophysical applications @xcite . in this paper
, we apply the backward finite - time lyapunov exponents ( b - ftle ) to study the attracting lcs of the photospheric turbulence .
we show that , like the `` corks '' , the b - ftle is able to trace the patterns of the magnetic network .
moreover , it gives a quantitative measure of the local rate of contraction / expansion of the photospheric flow .
our study is based on a 12 h high - resolution sequence of velocity images ( cadence @xmath9 ) of photospheric flow derived from the hinode - sot magnetograms , from 12-dec-2006 14:24 ut to 13-dec-2006 02:24 ut , by selecting a unipolar plage area of size 12.4 mm @xmath5 12.4 mm ( @xmath10 ) near the active region ar 10930 ( yeates et al .
the horizontal velocity field is a `` proxy '' extracted from the line - of - sight magnetic field ( @xmath11 ) using the fourier local correlation tracking method ( flct ) ( welsch et al .
2004 , 2012 ) .
it is worth pointing out that there are uncertainties in the correlation tracking that do not fully resolve the velocity fields in either space or time .
the detailed description of the data reduction procedure is given in @xcite and @xcite .
the flct method involves a number of optimum parameters which are determined by an autocorrelation analysis in order to maximize frame - to - frame correlations and ensure robustness in the velocity estimate . in view of the reduction of noise by averaging in the tracking procedure
, the actual velocity patterns of the photospheric flow might be more complicated than the `` proxy '' .
@xcite showed that the build - up of magnetic gradients in solar corona can be inferred directly from the photospheric velocity data , without recourse to magnetic field extrapolation .
they established the correspondence of the network of quasi - separatrix layers @xcite to the repelling lcs in the photospheric velocity . in this paper , we extend the analysis of @xcite to establish the correspondence of the network of high magnetic flux concentration to the attracting lcs in the photospheric velocity .
eulerian coherent structures can be extracted from the velocity field by computing the q - criterion , also known as the okubo - weiss parameter in 2d turbulence @xcite , @xmath12 , \label{eq q}\ ] ] where the frobenius matrix norm is adopted and @xmath13 and @xmath14 are defined by the decompositions of the gradient tensor of the velocity field @xmath15 and @xmath16 , \quad \omega = \frac{1}{2}[\nabla u - ( \nabla u)^{\top}],\ ] ] where @xmath17 denotes the transpose of a tensor . the strain - rate ( or deformation ) tensor @xmath13 is the symmetric part of @xmath18 ; its eigenvectors form an orthonormal set along the principle directions of fluid element deformation caused by tension , and its eigenvalues are real quantities that measure the rates of deformation in the corresponding directions .
the vorticity tensor @xmath14 is the anti - symmetric part of @xmath18 .
for example , the deformation calculated using the ulysses data shows that the magnetic tension causes local stretching of solar wind plasma elements mostly along the direction of the average magnetic field , whereas the solar wind vorticity tends to become perpendicular to the mean radial direction at large heliospheric distances @xcite .
snapshot images of eulerian coherent structures of the horizontal velocity field detected by hinode in a plage region near ar 10930 @xcite are shown in figure [ fig euler ] for four instances : 12-dec-2006 14:24 ut , 12-dec-2006 18:24 ut , 12-dec-2006 22:24 ut , 13-dec-2006 02:24 ut .
the first column shows magnetograms ; the second column depicts the streamlines computed using the method of line integral convolution , which displays the integral curves of @xmath19 in different tones of orange ; some vortex coherent structures ( patches ) can be visualized in these plots .
the third column depicts the @xmath20 plots of the deformation eulerian coherent structures , which are identified by the localized patches with high values of @xmath21 .
the fourth column depicts the @xmath20 plots of the vortical eulerian coherent structures , which are identified by the localized patches with high values of @xmath22 .
the average spatial size and lifetime of these vortices are similar to the magnetic bright points observed by @xcite .
the fifth column depicts the @xmath20 plots of the q - criterion given by eq .
( 1 ) , which partitions the turbulent flow into the deformation - dominated ( @xmath23 , blue patches ) and vorticity - dominated ( @xmath24 , red patches ) regions . the intensity of a turbulent flow is characterized by the magnitude of the root - mean - squared velocity fluctuations , deformation , and vorticity @xcite . the probability distribution functions ( pdf ) of eulerian coherent structures , associated with figure [ fig euler ] , are shown in figure [ fig pdfeuler ] .
some trends are evident
. figures [ fig pdfeuler](a ) and [ fig pdfeuler](b ) show that there are higher ( lower ) statistics for the low ( high ) values of the magnitudes of deformation and vorticity .
figure [ fig euler](c ) shows that for all four instances the q - criterion is distributed amongst negative and positive values , with non - gaussian pdf .
note that the pdf of the q - criterion shows a strong asymmetry skewed towards negative values , which indicates that in the plage region under study the photospheric turbulent flow is dominated by the strain - rate field .
this situation is similar to a laboratory experiment of the turbulence - flame interaction , where the flame surface straining is dominated by the deformation whereas the role of the vorticity is to curve the flame surface creating wrinkles @xcite .
the mean values of eulerian coherent structures for four instances are summarized in table 1 .
it confirms the features of pdf seen in figure [ fig pdfeuler ] , i.e. , in the plage region under study , on average : ( i ) the deformation dominates over the vorticity , ( ii ) the q - criterion shows negative values implying the dominance of the deformation coherent structures over the vortical coherent structures .
.mean values of the eulerian coherent structures [ cols="^,^,^,^",options="header " , ] although the deformation and vortical coherent structures can be extracted by the eulerian technique using the q - criterion , it relies on instantaneous snapshots of the velocity field . in the next section we show that the transport barriers in the turbulent flow can be detected with greater precision by the technique of lcs .
analogous to `` corks '' discussed in section 1 , the finite - time lyapunov exponent ( ftle ) can be computed by advecting a dense grid of tracer particles over the domain of interest .
consider a passive particle advected by the velocity field @xmath25 from an initial time @xmath26 .
we use the horizontal velocity field derived from the hinode sot data for the plage region of figure [ fig euler ] to solve the particle advection equation @xmath27 over a grid of initial positions @xmath28 until the final positions @xmath29 are reached after a finite integration time @xmath30 .
the particle trajectories are obtained by a fourth - order runge - kutta integrator with cubic spline interpolation in both space and time .
the finite - time lyapunov exponents of the particle trajectories for a 2d flow are calculated at each initial position @xmath28 as @xcite @xmath31 where @xmath32 ( @xmath33 ) are the eigenvalues of the finite - time right cauchy - green deformation tensor @xmath34 , @xmath35 is the deformation gradient , @xmath17 denotes the transpose and @xmath36 is the flow map for eq .
@xmath37 . in forward time
@xmath38 ) , the maximum ftle , @xmath39 , gives the finite - time average of the maximum rate of either divergence ( if @xmath40 ) or convergence ( if @xmath41 ) between the trajectories of a fiducial particle at @xmath42 and its neighboring particles .
the maximum stretching is found when the neighboring particle @xmath43 is such that @xmath44 is initially aligned with the eigenvector of @xmath45 associated with @xmath46 .
if two points are initially separated by a small distance @xmath47 at time @xmath26 , then their maximum separation at a later time @xmath48 will be @xmath49|\xi_0|$ ] .
@xmath50 provides information about stretching / contraction in another direction and can be useful to interpret the local dynamics of the fluid .
local minima in the maximum forward finite - time lyapunov exponent ( f - ftle ) provides a way to detect the position of the center of vortices in the velocity field since vortices may be viewed as material tubes of low particle dispersion . in backward time
@xmath51 ) , @xmath40 represents regions of the flow with convergence and @xmath41 , regions with divergence of nearby trajectories .
advecting a particle forward in time reveals the repelling lcs in the f - ftle field , which are the source of stretching in the flow , whereas advecting a particle backward in time reveals the attracting lcs in the b - ftle field along which particles congregate to form the observable patterns @xcite .
the properties of an lcs depend on the choice of the integration time @xmath30 , which should be chosen long enough for dominant features to be revealed , yet short enough for the ftle to be representative of the transient dynamics of interest .
three examples of repelling lcs for the phostospheric turbulent flow in the plage region under study are given in figures [ fig lcs](a)-(c ) , which show the maximum f - ftle @xmath39 computed by solving the particle advection equation ( 4 ) from the initial time @xmath26 = 12-dec-2006 14:24 ut to the final time @xmath52 , with @xmath53h ( figure [ fig lcs](a ) ) , @xmath54h ( figure [ fig lcs](b ) ) , and @xmath55h ( figure [ fig lcs](c ) ) , respectively .
thin ridges of large ( positive ) f - ftle in figures [ fig lcs](a)-(c ) represent the locally strongest repelling material lines in the photospheric flow , which confirm the results of @xcite using @xmath56h and + 12h .
as we increase integration time we notice some structures persist and become more sharply defined .
these structures can be considered most influential over time .
the structures that fail to persist as integration time is increased can be considered to be due to more transient flow features .
three examples of attracting lcs for the photospheric turbulent flow in the plage region under study are given in figures [ fig lcs](d)-(f ) , which show the maximum b - ftle @xmath39 computed by solving the particle advection equation ( 4 ) from the initial time @xmath26 = 13-dec-2006 02:24 ut to the final time @xmath52 , with @xmath57h ( figure [ fig lcs](d ) ) , @xmath58h ( figure [ fig lcs](e ) ) , and @xmath59h ( figure [ fig lcs](f ) ) , respectively .
thin ridges of large ( positive ) b - ftle in figures [ fig lcs](d)-(f ) represent the locally strongest attracting material lines in the photospheric flow . in order to illustrate the temporal evolution of lcs , we compute the forward- and backward - time ftle of the photospheric velocity field at @xmath26 = 12-dec-2006 18:24 ut and @xmath26 = 12-dec-2006 22:24 ut , using the same integration time @xmath60h .
figure [ fig lcs2 ] displays the repelling ( green ) and attracting ( red ) lcs for two @xmath26 , which form the `` lagrangian skeletons '' of the photospheric turbulence . attracting and repelling lcs act as barriers to particle transport .
a material line is a smooth curve of fluid particles advected by the velocity field @xcite . these attracting and repelling material lines are the analogous of stable and unstable manifolds of time - independent fields @xcite .
numerical studies of 2d flows have elucidated the role of material lines @xcite .
consider a steady flow , where the velocity field does not change with time . in the presence of counter - rotating vortices , hyperbolic ( saddle )
points are expected to be found .
the trajectories of passive scalars follow the velocity vectors in the vicinity of the hyperbolic point .
thus , particles lying on the stable manifold are attracted to the saddle point in the forward - time dynamics and trajectories on the unstable manifold converge to the saddle point in the backward - time dynamics @xcite .
two particles are said to straddle a manifold if the line segment connecting them crosses the manifold .
the maximum ftle typically has high values on the stable manifold in forward - time , since nearby trajectories straddling the manifold will experience exponential divergence when they approach the saddle point . similarly , the ftle field exhibits a local maximizing curve ( ridge ) along the unstable manifold in backward - time dynamics , since trajectories straddling the unstable manifold diverge exponentially when they approach the saddle point in reversed - time .
thus , ridges in the forward - time ftle field mark the stable manifolds of hyperbolic points and ridges in the backward - time ftle field mark the unstable manifolds .
analogously , for a time - dependent velocity field , regions of maximum material stretching generate ridges in the ftle field .
thus , the repelling material lines ( finite - time stable manifolds ) produce ridges in the maximum ftle field in the forward - time system and attracting material lines ( finite - time unstable manifolds ) produce ridges in the backward - time system , as seen in figures [ fig lcs ] and [ fig lcs2 ] .
in this paper , we showed that eulerian coherent structures and lagrangian coherent structures give very different information of the dynamics and structure of astrophysical plasma flows .
eulerian coherent structures give instantaneous information of plasma dynamics and structure at a given time , whereas lagrangian coherent structures account for the integrated effect of plasma dynamics and structure in a finite - time interval .
arguably , instantaneous quantities may be considered inappropriate to understand inherently transient phenomenon since such quantities may not properly convey the integrated behavior of constantly changing flow .
lcs typically account for such integrated behavior more naturally by considering the integrated fluid motion to reveal organizing flow features . in the context of photospheric turbulence , we showed in figure [ fig euler ] that the eulerian analysis is able to differentiate two types of coherent structures arising from either deformation or vorticity .
an enlargement of the rectangle region indicated at figure [ fig euler ] at 12-dec-2006 18:24 ut is depicted in figure [ fig zoomeuler2d ] .
the pattern of streamlines in figure [ fig zoomeuler2d](a ) identifies the localized region of a vortex in the photospheric flow .
figures [ fig zoomeuler2d](c ) and [ fig zoomeuler2d](d ) show that this vortex region is characterized by high values of the modulus of both deformation and vorticity .
figure [ fig zoomeuler2d](b ) shows that the q - criterion can differentiate clearly two subregions : one being the deformation eulerian coherent structures ( blue subregion ) and the other being the vortical eulerian coherent structures ( red subregion ) . for the sake of clarity , a corresponding 3d view of figure [ fig zoomeuler2d ]
is presented in figure [ fig zoomeuler3d ] .
our eulerian analysis shows that , on average , in the plage region the deformation eulerian coherent structures dominate over the vortical eulerian coherent structures in the turbulent photospheric flows , as seen in figure [ fig pdfeuler ] .
the lagrangian analysis is able to provide further information of magnetic and velocity fields of the photospheric turbulence missing in the eulerian analysis .
yeates et al .
( 2012 ) demonstrated that it is possible to make the link between the squashing q - factor and the maximum lyapunov exponent @xmath39 of the repelling lagrangian coherent structures , since both quantities are similar measures of the local rate of stretching at a given point and defined by the norm of the cauchy - green tensor of the deformation imposed by the field - line mapping .
note , however , that there are some differences between these two quantities .
while the squashing q - factor uses the frobenius norm of the tensor , @xmath61 uses the spectral norm ; while the squashing q - factor is dimensionless , @xmath61 has units of inverse time and includes the logarithm in its definition in eq .
the close similarity of these two quantities is shown in figure [ fig squash2d ] , where the squashing q - factor calculated for the same region of the photosphere under study is shown in figure [ fig squash2d](a ) and the corresponding repelling lagrangian coherent structures are shown in figure [ fig squash2d](b ) . both figures [ fig squash2d](a ) and [ fig squash2d](b ) are computed for @xmath26 = 12-dec-2006 14:24 ut and @xmath30 = + 12 h , the same as figure [ fig lcs](c ) .
figure [ fig squash2d](a ) adopts the logarithmic scaling since certain trajectories typically become exponentially separated in time , which shows the emergence of thin ridges of high values of the squashing q - factor representing the quasi - separatrix layers .
these ridges are interspersed by regions of low values of the squashing q - factor .
figure [ fig squash2d](b ) shows the emergence of thin ridges of high positive values of @xmath39 , representing the locally strongest repelling material surfaces in the photospheric flow .
these ridges are interspersed by regions of negative @xmath39 that represent converging trajectories of the plasma transport . evidently , the ridges of the quasi - separatrix layers in figure [ fig squash2d](a ) are co - spatial with the ridges of the photospheric flow in figure [ fig squash2d](b ) . for the sake of clarity ,
a 3d view of the rectangle regions of figure [ fig squash2d ] is presented in figure [ fig squash3d ] .
note that the spiky features seen in figure [ fig squash3d](a ) may be related to numerical issues which can be smoothed by applying a refined numerical procedure @xcite . as mentioned in section 1
, the lagrangian analysis based on corks can provide insight into the relationship between magnetic fields and plasma flows in the photosphere .
roudier et al .
( 2009 ) used a 48 h time sequence of the horizontal velocity data from the solar optical telescope ( sot ) onboard hinode to study the interactions between granular to supergranular scales in a quiet region of the photosphere .
they showed that the tree of fragmenting granules plays a crucial role in the advection of the magnetic field and in the build - up of the magnetic network .
in particular , the long time sequence of their analysis shows that the trajectory of plasma flows traced by corks matches exactly the position of the highest magnetic flux concentration of the network , which lies on supergranule boundaries .
our technique of lagrangian coherent structures is closely related to corks , but gives additional important information missing in the cork - based studies as demonstrated below .
figure [ fig blcs2d ] shows a comparison of the line - of - sight magnetic field @xmath11 measured by hinode / sot at @xmath62 = 13-dec-2006 02:24 ut ( figure [ fig blcs2d](a ) , same as figure [ fig euler ] ) and the attracting lagrangian coherent structures computed for @xmath26 = 13-dec-2006 02:24 ut and @xmath30 = -12 h ( figure [ fig blcs2d](b ) , same as figure [ fig lcs](f ) ) .
it is possible to identify similar patterns in figures [ fig blcs2d](a ) and [ fig blcs2d](b ) .
for example , a superposition of the rectangle regions marked in figures [ fig blcs2d](a)-(b ) is shown in figure [ fig blcs2d](c ) ( where the thresholded b - ftle with @xmath63 is plotted ) that shows clearly the proximity of the trajectory of plasma flows to the magnetic field , similar to the cork results of roudier et al .
in addition , figure [ fig blcs2d](b ) shows the emergence of thin ridges of high positive values of @xmath39 , representing the locally strongest attracting material surfaces in the photospheric flow .
these ridges are interspersed by regions of negative @xmath39 that represent diverging trajectories of the plasma transport . for the sake of clarity , a 3d view of the rectangle regions of figure [ fig blcs2d ]
is presented in figure [ fig blcs3d ] . in order to confirm the validity of figures [ fig blcs2d ] and [ fig blcs3d ]
, we performed 3d numerical simulations of large - scale dynamos in turbulent compressible convection with uniform horizontal shear and rotation .
the dimensionless compressible magnetohydrodynamics equations are solved in a simulation box divided into three layers , an upper cooling layer , a convectively unstable layer , and a stable overshoot layer , with constant gravity in the vertical direction , as described by @xcite .
the box has dimensions @xmath64 , where @xmath65 is the depth of the convectively unstable layer , and shearing periodic boundary conditions are used in the horizontal direction . in the vertical direction we use stress - free boundary conditions for the velocity field and vertical field conditions for the magnetic field . in our simulations we adopt dimensionless quantities , setting @xmath66 , where @xmath67 is gravity , distance is in units of @xmath65 , density in units of the initial value at the base of convective layer @xmath68 , time in units of the free fall time @xmath69 , velocity in units of @xmath70 and magnetic field in units of @xmath71 . the physical parameters ( e.g. kinematic viscosity , resistivity , heat conductivity , rotation velocity , etc .
) are chosen to ensure the onset of a large - scale magnetic field in a moderately turbulent velocity field . for a detailed setup ,
we direct the reader to run d2 in @xcite .
the model is solved with the pencil code , which employes sixth - order finite - differences in space and third - order variable step runge - kutta integration in time .
figure [ fig blcs_sim2d](a ) shows a 2d image of the vertical component of the magnetic field @xmath11 near the top of the convective layer at a time @xmath26 when the magnetic energy growth has already saturated .
this figure shows that the simulated large - scale turbulent dynamo reproduces the pattern of convective cells , similar to granulations in the photosphere .
in particular , it shows that the network of high concentration of magnetic flux is located in the boundaries between the convective cells , corresponding to the intergranular lanes .
the elongated convective cells in figure [ fig blcs_sim2d ] result from the shear .
since there is an imposed large - scale flow in the @xmath72 direction , it is expected that the convective cells would be elongated along this direction .
this is more pronounced after a strong large - scale magnetic field develops due to the dynamo mechanim .
figure [ fig blcs_sim2d](b ) shows the b - ftle computed for the horizontal velocity field @xmath73 at @xmath26 using @xmath74 time units .
the ridges of high values of b - ftle in figure [ fig blcs_sim2d](b ) indicate the locations of converging plasma flows . in figure [ fig blcs_sim2d](c )
we plot the superposition of the thresholded b - ftle ( with @xmath75 ) on top of @xmath11 . for the sake of clarity
, a 3d view of the rectangle regions of figure [ fig blcs_sim2d ] is given in figure [ fig blcs_sim3d ] , where the modulus of @xmath11 is plotted in comparison with the b - ftle .
evidently , figures [ fig blcs_sim2d ] and [ fig blcs_sim3d ] render support for the results of figures [ fig blcs2d ] and [ fig blcs3d ] that there is correspondence of the network of high magnetic flux concentration to the attracting lcs in the photospheric velocity . in conclusion
, we extended the observational study by @xcite of lagrangian coherent structures in the photospheric velocity to show that in addition to the correspondence of the network of quasi - separatrix layers in the repelling lcs demonstrated by @xcite , there is a correspondence of the network of high magnetic flux concentration to the attracting lcs in the photospheric flows , confirmed by both observations and numerical simulations .
hence , the repelling and attracting lcs of the photospheric velocity provide complementary informations of the local dynamics and topology of solar magnetic fields .
although the corks coincide with the network of magnetic flux , they are not the best way to find the transport barriers and the directions of the flow convergence .
we showed that the attracting lcs are more suitable for these tasks because they trace continuous curves ( material lines ) that are the explicit structures potentially responsible for observed physical patterns and at the same time measure the local rate of contraction / expansion in the flow .
while eulerian coherent structures provide useful insights of the snapshot of the photospheric flows , they fail to reveal the fine structures of the transport barriers and the local dynamics of the flows that are readily provided by the repelling and attracting lagrangian coherent structures .
a.c.l.c . acknowledges the support of cnpq , the award of a marie curie fellowship by the european commission , and the hospitality of paris observatory .
elr acknowledges the financial support of cnpq ( brazil ) and fapesp ( brazil ) .
we thank petri kpyl for providing the configuration files for the pencil code mhd simulations .
cadavid , a. c. , lawrence , j. k. , christian , d. j. , jess , d. b. , & mathioudakis , m. 2012 , in `` magnetic fields from the photosphere to the corona '' , t. r. rimmele , m. collados vera , et al .
( eds . ) , asp conf .
series , vol .
75 - 82 * pdf of eulerian coherent structures ( ecs ) as a function of : ( a ) * @xmath21 , * ( b ) * @xmath22 , * ( c ) * q - criterion ; at 12-dec-2006 14:24 ut ( green ) , 12-dec-2006 18:24 ut ( yellow ) , 12-dec-2006 22:24 ut ( red ) , 13-dec-2006 02:24 ut ( black ) . ]
= 13-dec-2006 02:24 ut and * ( b ) * the b - ftle for @xmath26 = 13-dec-2006 02:24 ut and @xmath30 = -12 h. a superposition of the rectangle regions of ( a ) and ( b ) ( green ) is shown in * ( c ) * , where the thresholded @xmath63 is applied to b - ftle . ] | we study coherent structures in solar photospheric flows in a plage in the vicinity of the active region ar 10930 using the horizontal velocity data derived from hinode / sot magnetograms .
eulerian and lagrangian coherent structures are detected by computing the q - criterion and the finite - time lyapunov exponents of the velocity field , respectively .
our analysis indicates that , on average , the deformation eulerian coherent structures dominate over the vortical eulerian coherent structures in the plage region .
we demonstrate the correspondence of the network of high magnetic flux concentration to the attracting lagrangian coherent structures ( a - lcs ) in the photospheric velocity based on both observations and numerical simulations .
in addition , the computation of a - lcs provides a measure of the local rate of contraction / expansion of the flow . |
though shannon showed in 1956 @xcite that noiseless feedback does not increase the capacity of memoryless channels , feedback s other benefits have made it a staple in modern communication systems .
feedback can simplify the encoding and decoding operations and has been incorporated into incremental redundancy ( ir ) schemes proposed as early as 1974 @xcite .
hagenauer s work on rate - compatible punctured convolutional ( rcpc ) codes allows the same encoder to be used in various channel conditions and uses feedback to determine when to send additional coded bits @xcite .
the combination of ir and hybrid arq ( harq ) continues to receive attention in the literature @xcite and industry standards such as 3gpp .
although it can not increase capacity in point - to - point channels , the information - theoretic benefit of feedback for reducing latency through a significant improvement in the error exponent has been well understood for some time .
( see , for example , @xcite . )
recent work @xcite casts the latency benefit of feedback in terms of block length rather than error exponent , generating new interest in the practical value of feedback for approaching capacity with a short average block length .
polyanskiy et al . provided bounds for the maximum rate that can be accomplished with feedback for a finite block length @xcite and also demonstrated the energy - efficiency gains made possible by feedback @xcite . in the variable - length feedback with termination ( vlft ) scheme
, @xcite uses an elegant , single , `` stop feedback '' symbol ( that can occur after any transmitted symbol ) that facilitates the application of martingale theory to capture the essence of how feedback can allow a variable - length code to approach capacity .
a compelling example from @xcite shows that for a binary symmetric channel with capacity 1/2 , the average block length required to achieve 90% of the capacity is smaller than 200 symbols . for practical systems such as hybrid arq ,
the `` stop feedback '' symbol may only be feasible at certain symbol times because these systems group symbols together for transmission in packets , so that the entire packet is either transmitted or not . in @xcite , chen et al .
used a code - independent rate - compatible sphere - packing ( rcsp ) analysis to quantify the latency benefits of feedback in the context of such grouped transmissions .
chen et al . focused on the awgn channel and also showed that capacity can be approached with surprisingly small block lengths , similar to the results of @xcite . using the rcsp approach of chen et al .
as its foundation , this paper introduces an optimization technique and uses it to explore how closely one may approach capacity with only a handful of incremental transmissions . for a fixed number of information bits @xmath0 and a fixed number of maximum transmissions
@xmath1 before giving up to try again from scratch , a numerical optimization determines the block lengths of each incremental transmission to maximize the expected throughput .
we consider only @xmath1@xmath2@xmath3 and show that this is sufficient to achieve more than 90% of capacity while requiring surprisingly small block lengths similar to those achieved by polyanskiy et al . and
chen et al .
while rcsp is an idealized scheme , it provides meaningful guidance for the selection of block lengths and the sequence of target decoding error rates , which we call the decoding error trajectory . a 1024-state rate - compatible
punctured tail - biting convolutional code using the block lengths determined by our rcsp optimization technique achieves the rcsp decoding error trajectory and essentially matches the throughput and latency performance of rcsp for @xmath1@xmath4@xmath5 transmissions .
our results , like those of polyanskiy et al . and chen et al . , assume that the receiver is able to recognize when it has successfully decoded . the additional overhead of , for example ,
a cyclic redundancy check ( crc ) has not been included in the analysis .
longer block lengths would be required to overcome this overhead penalty .
the paper is organized as follows : section [ sec : analysis ] reviews the rcsp analysis .
section [ sec : algo ] describes the rcsp numerical optimization used to determine transmission lengths and shows the throughput vs. latency performance achieved by using these transmission lengths for up to six rate - compatible transmissions .
this performance is compared with a version of vlft scheme proposed by polyanskiy et al .
section [ sec : sims ] introduces the decoding error trajectory and shows how rcsp performance can be matched by a real convolutional code using the transmission lengths identified in the previous section .
section [ sec : outage ] shows how the rcsp analysis can be applied to scenarios that involve strict latency and outage probability constraints .
section [ sec : conc ] concludes the paper .
to review the sphere - packing analysis presented in @xcite for a memoryless awgn channel , consider a codebook of size @xmath6 that maps @xmath7 information symbols into a length-@xmath8 codeword with rate @xmath9 .
the channel input and output can be written as : @xmath10 where @xmath11 is the output ( received word ) , @xmath12 is the codeword of the @xmath13th message , and z is an @xmath8-dimensional i.i.d .
gaussian vector .
let the received snr be @xmath14 and assume without loss of generality that each noise sample has unit variance .
the average power of received word @xmath11 is then @xmath15 . as in @xcite , the largest possible squared decoding radius @xmath16 assuming that the decoding spheres occupy all available volume is @xmath17 a bounded - distance decoder declares any message within a distance @xmath18 of codeword @xmath12 to be message @xmath13
otherwise , a decoding error is declared . because the sum of the squares of the @xmath8 gaussian noise samples obeys a chi - square distribution with @xmath8 degrees of freedom ,
the probability @xmath19 of decoding error associated with decoding radius @xmath18 is @xmath20 where the @xmath21 are standard normal distributed random variables with zero mean and unit variance and @xmath22 is the cdf of a chi - square distribution with @xmath8 degrees of freedom .
the idea of rcsp is to assume that sphere - packing performance can be achieved by each transmission in a sequence of rate - compatible transmissions .
thus the idealized sphere - packing analysis is applied to a modified incremental redundancy with feedback ( mirf ) scheme as described in @xcite .
mirf works as follows : @xmath0 information symbols are coded with an initial block length @xmath23 . if the receiver can not successfully decode , the transmitter will receive a nack and send @xmath24 extra symbols . the decoder attempts to decode again using all received symbols for the current codeword , i.e. , with block length @xmath25 .
the process continues for @xmath26 .
the decoded block length @xmath27 is @xmath28 and the code rate is @xmath29 .
if decoding is not successful after @xmath1 transmissions , the decoder discards the @xmath1 transmissions and the process begins again with the transmitter resending the @xmath30 initial symbols .
this scheme with @xmath1=@xmath31 is standard arq .
the squared decoding radius of the @xmath13th cumulative transmission is @xmath32 and the marginal probability of decoding error @xmath33 associated with decoding radius @xmath34 is @xmath35 where the @xmath36 are standard normal distributed random variables with zero mean and unit variance .
however , this marginal probability is not what is needed .
the probability of a decoding error in the @xmath13th transmission depends on previous error events .
indeed , conditioning on previous decoding errors @xmath37 makes the error event @xmath38 more likely than the marginal distribution would suggest .
the joint probability @xmath39 is @xmath40 we compute the expected number of channel uses ( i.e. , latency or average block length ) @xmath41 by summing the incremental transmission lengths @xmath42 weighted by the probability of error in the prior cumulative transmission and dividing by the probability of success by the last ( @xmath1th ) transmission ( as in arq ) , according to @xmath43 this expression does not consider delay due to decoding operations .
the corresponding throughput @xmath44 is given by @xmath45
in the special case of @xmath1=@xmath31 ( when only the initial transmission of length @xmath30 is ever transmitted ) , mirf is arq . in this case
the expected number of channel uses given by ( [ eqn : lambda_tau ] ) can be simplified as follows ( with @xmath47 ) : @xmath48 which yields an expected throughput of @xmath49 if we fix the number of information bits @xmath0 , ( [ eqn : rt_arq ] ) becomes a quasiconcave function of the initial code rate @xmath50 , allowing the optimal code rate @xmath51 , which maximizes the throughput @xmath44 for a given @xmath0 , to be found numerically @xcite .
[ fig : latvthroughput_a ] plots the maximum achievable throughput in the @xmath1@xmath4@xmath31 ( arq ) rcsp scheme as the red ( diamond markers ) curve . in @xcite , chen et al
. demonstrated one specific rcsp scheme with ten transmissions that could approach capacity with low latency .
specifically , the transmission lengths were fixed to @xmath30@xmath4@xmath53 and @xmath54@xmath4@xmath55 , while @xmath0 was varied to maximize throughput .
this paper builds on the intuition of the arq case presented above .
both @xmath0 and the number of transmissions @xmath1 are fixed , and a search identifies the set of transmission lengths @xmath56 that maximizes throughput . we seek to identify approximately how much throughput can be achieved using feedback with a small number of incremental transmissions , specifically @xmath1@xmath2@xmath3 .
furthermore , we seek insight into what the transmission lengths should be and what decoding error rates allow the sequence of transmissions to be most efficient . for @xmath1@xmath57@xmath31 , identifying the transmission lengths
@xmath56 which minimize the latency @xmath41 in ( [ eqn : lambda_tau ] ) is not straightforward due to the joint decoding error probabilities in ( [ eqn : pzeta_i ] )
. however , the restriction to a small @xmath1 allows exact computation of in mathematica , avoiding the approximations of@xcite . to reflect practical constraints ,
we restrict the lengths @xmath42 to be integers .
the computational complexity of , which increases with the transmission index @xmath13 , forces us to limit attention to a well - chosen subset of possible transmission lengths .
thus , our present results may be considered as lower bounds to what is possible with a fully exhaustive optimization .
[ fig : latvthroughput_a ] shows the throughput vs. latency performance achieved by rcsp for @xmath58 on an awgn channel with snr 2.0 db . as @xmath1 is increased , each additional
retransmission brings the expected throughput @xmath44 closer to the channel capacity , though with diminishing returns .
the points on each curve in fig .
[ fig : latvthroughput_a ] represent values of @xmath0 ranging from 16 to 256 information bits .
[ fig : latvthroughput_a ] shows , for example , that by allowing up to four retransmissions ( @xmath1@xmath4@xmath5 ) with @xmath0@xmath4@xmath53 , rcsp can achieve 91% of capacity with an average block length of 102 symbols .
similar results are obtained for other snrs .
vlft achievability results for the awgn channel based on @xcite are shown in fig .
[ fig : latvthroughput_b ] for comparison .
both the original vlft scheme , in which the transmission may be ended after any symbol , and a constrained version of vlft using the same block lengths and feedback structure as @xmath1=@xmath5 rcsp are presented .
the original vlft closely approaches capacity with a latency on the order of 200 symbols .
rcsp is unable to match vlft because the overall rcsp transmission can only be terminated after one of the @xmath1 incremental transmissions completes .
if vlft is constrained in the same way , its performance is initially worse than rcsp because random coding does not achieve ideal sphere packing with short block lengths . at an average latency of 200 , constrained
vlft performance becomes similar to the comparable rcsp scheme .
the vlft achievability curve evaluates ( * ? ? ? * theorem 10 ) using the upper bound of ( 162 ) with i.i.d .
gaussian inputs with average power equal to the power constraint @xmath14
. such codebooks will sometimes violate the 2 db power constraint . to address this ,
the average power should be slightly reduced and codebooks violating the power constraint should be purged , which will lead to a small performance degradation .
alternatively , codebooks or even codewords can be constrained to meet the power constraint with equality .
further analysis of vlft codes more carefully considering the power constraint for the awgn channel will be the subject of future work .
rcsp makes the rather optimistic assumption that a family of rate - compatible codes can be found that performs , at each rate , equally well as codes that pack decoding spheres so well that they use all of the available volume .
a variety of well - known upper bounds on the packing density @xmath59 indicate that the maximum packing density decreases as the dimension @xmath60 increases ( e.g. , @xmath61 ) @xcite , making such codes difficult to find .
however , we show in this section that a rate - compatible tail - biting convolutional code can indeed match the performance of rcsp , at least for @xmath1@xmath4@xmath5 .
we consider two rate @xmath31/@xmath62 convolutional codes from @xcite : a 64-state code with generator polynomial ( @xmath63,@xmath64,@xmath65)=(@xmath66 ) and a 1024-state code with ( @xmath63,@xmath64,@xmath65)=(@xmath67 ) , where the generator notation is octal .
high rate codewords are created by pseudorandom rate - compatible puncturing of the rate @xmath31/@xmath62 mother codes .
we restrict our attention to tail - biting implementations of these convolutional codes because the throughput efficiency advantage is important for the relatively small block lengths we consider .
simulations compare the performance of these two codes in the mirf setting for the awgn channel with snr 2 db , as shown in fig .
[ fig : latvthroughput_b ] .
the simulations presented here focus on the @xmath0@xmath4@xmath53 case .
.optimal rcsp transmission lengths for @xmath68 and snr 2 db . [ cols="^,^,^,^,^,^",options="header " , ] [ table : m5_steps ] the transmission lengths @xmath42 used in the simulations are those identified by the rcsp optimization . table [ table : m5_steps ] shows the results of the @xmath1@xmath4@xmath5 optimization ( i.e. , the set of lengths @xmath42 found to achieve the highest throughput ) .
thus our simulations used @xmath30@xmath4@xmath69 , @xmath24@xmath4@xmath70 , @xmath71@xmath4@xmath72 , @xmath73@xmath4@xmath70 , @xmath74@xmath4@xmath75 .
the induced code rates of the cumulative blocks are @xmath53/@xmath69=@xmath76 , @xmath53/@xmath77=@xmath78 , @xmath53/@xmath79=@xmath80 , @xmath53/@xmath81=@xmath82 and @xmath53/@xmath83=@xmath84 .
[ fig : rates_vs_k ] shows these rates as well as the rates for other values of @xmath0 according to the rcsp optimization for @xmath68 .
note that for every value of @xmath0 the initial code rate is above the channel capacity of 0.6851 .
this is the benefit of feedback : it allows the decoder to capitalize on favorable noise realizations by attempting to decode early , instead of needlessly sending additional symbols .
the rcsp optimization also computes the joint decoding error probabilities of , which we call the `` decoding error trajectory '' .
if we can find a rate - compatible family that achieves this decoding error trajectory , then we can match the rcsp performance .
[ fig : prerrorm5k64 ] shows the @xmath0=@xmath53 decoding error trajectories for the rcsp cases studied in figs . 1 and 2 ( shown as @xmath1 discrete points in fig .
[ fig : prerrorm5k64 ] for each value of @xmath85 ) and for constrained vlft for @xmath68 and 64-state and 1024-state convolutional code simulations for @xmath1=@xmath5 .
the dashed line represents the marginal probability of error for a sphere - packing codebook as in which was recognized in @xcite as a tight upper bound for the joint probabilities of error given by .
this tight upper bound can serve as a performance goal for practical rate - compatible code design across a wide range of block lengths .
while the 64-state code is not powerful enough to match rcsp performance , the 1024-state code closely follows the rcsp trajectory for @xmath68 . thus there
exist practical codes , at least in some cases , that achieve the idealized performance of rcsp . indeed , fig .
[ fig : latvthroughput_b ] plots the @xmath86 points of the two convolutional codes , demonstrating that the 1024-state code achieves 91% of capacity with an average latency of 102 symbols , almost exactly coinciding with the rcsp point for @xmath1@xmath4@xmath5 and @xmath0@xmath4@xmath53 .
the convolutional code s ability to match a mythical sphere - packing code is due to maximum likelihood ( ml ) decoding , which has decoding regions that completely fill the multidimensional space ( even in high dimensions ) .
these simulation results assume that the receiver is able to recognize when it has successfully decoded .
this same assumption is made by the rcsp analysis , the vlft scheme of polyanskiy et al .
, and the mirf scheme of chen et al .
while this assumption does not undermine the essence of this demonstration of the power of feedback , its practical and theoretical implications must be reviewed carefully , especially when very short block lengths are considered .
an important practical implication is that the additional overhead of a crc required to avoid undetected errors will drive real systems to somewhat longer block lengths than those presented here .
this will affect the choice of error control code .
an important implication is that this analysis can not be trusted if the block lengths become too small .
this assumption allows block errors to become block erasures at no cost .
consider the binary symmetric channel ( bsc ) : if the block length is allowed to shrink to a single bit , then this assumption turns the zero capacity bsc with transition probability @xmath87 into a binary erasure channel with probability @xmath87 , which has a capacity of @xmath87 instead of zero .
both the practical and theoretical problems of this assumption diminish as block length grows .
however , a quantitative understanding of the cost of knowing when decoding is successful and how that cost changes with block length is an important area for future work .
mirf has an outage probability of zero because it never stops trying until a message is decoded correctly .
with slight modifications , the mirf scheme and the rcsp transmission length optimization can incorporate strict constraints on latency ( so that the transmitter gives up after @xmath1 transmissions ) and outage probability ( which would then be nonzero ) . to handle these two new constraints ,
we restrict @xmath88 to be less than a specified @xmath89 . without modifying the computations of @xmath39 in ( [ eqn : pzeta_i ] )
, the optimization is adapted to pick the set of lengths that yields the maximum throughput s.t .
when there is a decoding error after the @xmath1th transmission , the transmitter declares an outage event and proceeds to encode the next @xmath0 information bits .
this scheme is suitable for delay - sensitive communications , in which data packets are not useful to the receiver after a deadline has passed .
the expected number of channel uses @xmath41 is now given by @xmath91 the expected throughput @xmath44 is again given by ( [ eqn : rt ] ) .
the purpose of this paper is to bring the information theory of feedback and the communication practice of feedback closer together . beginning with the idealized notion of rate - compatible codes with decoding spheres that completely fill the available volume
, the paper eventually demonstrates a convolutional code with performance strikingly similar to the ideal rate - compatible sphere - packing ( rcsp ) codes .
an optimization based on rcsp identifies the highest throughput possible for a fixed @xmath0 and @xmath1 .
this optimization provides the lengths of the initial and subsequent transmissions and the sequence of decoding error probabilities or `` decoding error trajectory '' that characterize the throughput - maximizing performance .
the rcsp decoding error trajectories computed in this paper are tightly bounded by the marginal error probability of sphere packing .
designing a code with a similar error trajectory will thus yield comparable latency performance .
rcsp predictions and simulation results agree in demonstrating that feedback permits 90% of capacity to be achieved with about 100 transmitted symbols assuming that the decoder knows when it has decoded correctly . however , the implications of this assumption for short block lengths warrant further investigation .
vlft performance shows that if the transmission could be stopped at any symbol ( rather than only at the end of each incremental transmission ) capacity is closely approached with an average latency of 200 symbols , but a more careful analysis of vlft in light of the awgn power constraint is warranted .
the authors would like to thank yury polyanskiy for helpful conversations regarding the vlft analysis . | recent work by polyanskiy et al . and chen et al . has excited new interest in using feedback to approach capacity with low latency .
polyanskiy showed that feedback identifying the first symbol at which decoding is successful allows capacity to be approached with surprisingly low latency .
this paper uses chen s rate - compatible sphere - packing ( rcsp ) analysis to study what happens when symbols must be transmitted in packets , as with a traditional hybrid arq system , and limited to relatively few ( six or fewer ) incremental transmissions .
numerical optimizations find the series of progressively growing cumulative block lengths that enable rcsp to approach capacity with the minimum possible latency .
rcsp analysis shows that five incremental transmissions are sufficient to achieve 92% of capacity with an average block length of fewer than 101 symbols on the awgn channel with snr of 2.0 db .
the rcsp analysis provides a decoding error trajectory that specifies the decoding error rate for each cumulative block length . though rcsp is an idealization , an example tail - biting convolutional code matches the rcsp decoding error trajectory and achieves 91% of capacity with an average block length of 102 symbols on the awgn channel with snr of 2.0 db .
we also show how rcsp analysis can be used in cases where packets have deadlines associated with them ( leading to an outage probability ) . |
let @xmath2 be an algebra and @xmath4 .
we say that @xmath1 _ absorbs _ @xmath2 , denoted by @xmath5 , iff there exists an idempotent term @xmath6 in @xmath2 ( that is , @xmath7 for each @xmath8 ) such that for each @xmath8 and @xmath9 we have @xmath10 the notion of absorption was developed a few years ago by barto and kozik , and immediately found many applications @xcite .
we would particularly like to mention that bulatov s dichotomy theorem for conservative csps @xcite , with a deep and complicated proof ( nearly 70 pages long ) , was reproved using these techniques on merely 10 pages @xcite .
loosely speaking , the main idea of absorption is that , when @xmath5 where @xmath1 is a proper subalgebra of @xmath2 , then some induction - like step can often be applied .
this naturally leads to the following question : given a finite algebra @xmath2 and its subalgebra @xmath1 , is it decidable whether @xmath5 ?
this question turns out to be quite hard .
let us mention that the notion of absorption emerged as a generalization of the notion of the so - called near - unanimity term ( in particular : an idempotent finite algebra @xmath2 has a near - unanimity term iff every singleton absorbs @xmath2 ) .
it was asked in 1995 whether the existence of a near - unanimity term in a finite algebra @xmath2 is decidable @xcite , and it took a while to finally prove that it is @xcite ( another interesting point here is that , before this proof appeared , there were some evidences suggesting that the answer is actually negative ) .
some recent results on deciding absorption are given in @xcite ; as expected , the proposed algorithms are quite complex . in this paper
we show that in semigroups the absorption is much easier to grasp .
namely , for absorption in semigroups , in theorem [ tt1 ] we provide a necessary and sufficient condition that is very easy to check .
after that , we turn to @xmath0-ary semigroups , that is , algebras @xmath11 where @xmath12 is an @xmath0-ary associative operation .
we conjecture an analogue of the necessary and sufficient condition for @xmath13 from theorem [ tt1 ] , and we then prove the conjecture in the following cases : when @xmath12 is commutative , when @xmath3 and when @xmath2 is an idempotent ternary semigroup .
let us say a few words on a possible application of these results .
namely , one of the most interesting algebraic results toward the csp dichotomy conjecture of feder and vardi @xcite is the proof that , if a finite relational structure @xmath14 does not admit any so - called weak near - unanimity ( wnu ) polymorphism , then @xmath15 is @xmath16-complete ( see @xcite , where bulatov , jeavons and krokhin gave a different algebraic sufficient condition for @xmath15 to be @xmath16-complete , and @xcite , where marti and mckenzie showed that this condition is equivalent to the nonexistence of a wnu polymorphism ) .
bulatov , jeavons and krokhin conjectured that the other direction also holds , that is , that the existence of a wnu polymorphism compatible with @xmath14 implies that @xmath15 is in @xmath17 ( this is known under the name algebraic dichotomy conjecture ) .
this has been checked for some relational structures of a special form , as well as for all relational structures but given the existence of a wnu polymorphism of a special form , and in all the cases known so far the results agree with the conjecture . in many of these
works the absorption was the key ingredient in the proof ( see , e.g. , the references from the beginning of this section ) . in particular , by the result of jeavons , cohen and gyssens @xcite , we know that whenever @xmath14 admits a semilattice polymorphism ( a semilattice operation is a binary operation that is idempotent , commutative and associative ) , then @xmath15 is in @xmath17 .
theorem [ tt1 ] from the present paper gives an exact description of when an algebra is absorbed by its subalgebra in the class of algebras with a binary associative operation ( which is a wider class than the class of algebras with a semilattice operation ) .
this provides a direct link between theorem [ tt1 ] and the current line of attack on the dichotomy conjecture .
concerning our generalization to @xmath0-ary semigroups , so far there is no result ( at least up to the author s knowledge ) toward the dichotomy conjecture that directly relates to conjecture [ conj ] in a similar manner ; however , since there are many results of this kind toward the dichotomy conjecture and many researchers are actively working on it , it is not hard to imagine that such a result exists and is just waiting to be discovered , and in fact , conjecture [ conj ] might serve as a motivation for it . and of course , speaking about the notion of absorption itself , theorem [ tt1 ] and conjecture [ conj ] may shed some light on the ( presently quite unclear ) behavior of absorption , since we now have a natural class of algebras in which the absorption behaves in a very predictable ( but nontrivial ) way .
it might be a very useful research direction to discover whether there is a deeper reason for this nice behavior of absorption in semigroups and ( conjecturally ) @xmath0-ary semigroups , and whether this reason may help to describe the behavior of absorption in other classes of algebras .
the main ( in fact , the only ) theorem in this section is the following one . [ tt1 ] let @xmath18 be a semigroup , and let @xmath19 .
then @xmath5 if and only if @xmath20 and @xmath21 for each @xmath8 , @xmath22 , and there exists a positive integer @xmath23 such that @xmath24 for each @xmath8 .
( @xmath25 ) : assume that the condition from the statement holds , and let us prove that @xmath5 .
choose a positive integer @xmath23 such that @xmath24 for each @xmath8 , and let @xmath26 .
for any @xmath8 we have @xmath27 , that is , @xmath6 is an idempotent term .
further , for any @xmath8 , @xmath22 , we have @xmath28 and @xmath29 , which proves that @xmath6 is an absorbing term .
( @xmath30 ) : let @xmath5 , and let @xmath6 be an absorbing term . since @xmath6 is an idempotent term , we trivially get that there exists a positive integer @xmath23 such that @xmath24 for each @xmath8 ( in particular , @xmath31 is the length of the term @xmath6 ) .
therefore , we are left to prove that @xmath20 and @xmath21 for each @xmath8 , @xmath22 .
let @xmath6 be a term in @xmath32 variables , which are named in such a way that the leftmost variable in @xmath33 is @xmath34 .
let @xmath35 denote the number of times the variable @xmath36 appears in @xmath33 .
let @xmath8 and @xmath22 be given .
we evaluate @xmath37 . because @xmath37 begins with @xmath38 and @xmath39
, we easily conclude @xmath40 since @xmath41 and @xmath6 is an absorbing term , by the previous equality we get @xmath42 let @xmath43 be any positive integer greater than @xmath44 such that @xmath45 . note that @xmath46 for @xmath47 , denote @xmath48 ( the expression @xmath49 is at the @xmath50 coordinate and at all the coordinates denoted by
@xmath51 " we put @xmath52 ) . since @xmath53 ( see ( [ abe ] ) ) , @xmath41 and @xmath6 is an absorbing term , we get @xmath54 as we have observed earlier , @xmath39 , and we thus conclude that @xmath55 evaluates to a power of @xmath38 . in particular , since @xmath56 appears @xmath57 times and @xmath49 appears @xmath35 times , we obtain @xmath58 consider the expression @xmath59 by ( [ abe ] ) we get @xmath60 which together with ( [ tib ] ) gives @xmath61 we further have @xmath62[c]{\scriptsize(\ref{tiizraz})}}{\approx}(ab)^{(r-(m-1)d_1)d_1+\sum_{i=2}^m(d_1 ^ 2+rd_i)}\\ & { } = ( ab)^{rd_1-(m-1)d_1 ^ 2+(m-1)d_1 ^ 2+\sum_{i=2}^mrd_i}\\ & { } = ( ab)^{r(d_1+d_2+\cdots+d_m)}=(ab)^{rk}. \end{array}\ ] ] since @xmath63 , it follows that @xmath64 whenever @xmath65 .
therefore , @xmath66 together with ( [ aaa ] ) , this gives @xmath20 , which was to be proved .
the proof that @xmath21 is analogous .
this completes the proof of theorem [ tt1 ] .
we say that an @xmath0-ary operation @xmath67 is _ associative _ iff @xmath68 for every @xmath69 .
an algebra @xmath11 , where @xmath12 is an @xmath0-ary associative operation , is called an _
@xmath0-ary semigroup_. instead of @xmath70 we shall often write @xmath71 , instead of the expressions from ( [ nasoc ] ) we shall write @xmath72 etc . however , we have to keep in mind that such an expression , say @xmath73 , is defined in @xmath2 if and only if @xmath74 . on the other hand ,
any expression of the form @xmath73 ( no matter whether @xmath74 or not ) will be called _ word_. even if the word @xmath75 is not defined in @xmath2 , we shall still write @xmath76 for the concatenation @xmath77 ( where @xmath75 is repeated @xmath78 times ) , but we need to be very careful not to apply any possible identities from @xmath2 on such a word ; for example , if @xmath2 is an idempotent ternary semigroup and @xmath79 , then @xmath80 is a valid way to write @xmath81 ( which is defined in @xmath2 ) , but we can not deduce @xmath82 .
the notation @xmath76 is defined also for @xmath83 , and in that case it stands for the empty word " , that is , @xmath84 means simply @xmath85 .
we believe that theorem [ tt1 ] can be generalized for @xmath0-ary semigroups in the following way .
[ conj ] let @xmath11 be an @xmath0-ary semigroup , and let @xmath19
. then the following conditions are equivalent : 1 . @xmath5 ; 2 .
@xmath86 and @xmath87 for each @xmath8 , @xmath22 , and there exists a positive integer @xmath23 such that @xmath24 for each @xmath8 ; 3 .
@xmath88 whenever at least one of @xmath89 belongs to @xmath90 , and there exists a positive integer @xmath23 such that @xmath24 for each @xmath8 .
we say that an @xmath0-ary operation @xmath12 is _ commutative _ iff @xmath91 for any @xmath89 and any permutation @xmath92 of the set @xmath93 .
we now prove conjecture [ conj ] in the case when @xmath12 is commutative , and then in two more cases , namely when @xmath3 and when @xmath2 is an idempotent ternary semigroup .
[ t1 ] conjecture [ conj ] holds when @xmath12 is commutative .
the implications @xmath94 and @xmath95 are easy , and in fact we shall not use the commutativity of @xmath12 in their proofs .
@xmath94 : let @xmath96 be given , and let @xmath97 for some @xmath98 . then @xmath99 which implies that it is enough to prove @xmath100 and @xmath101 . and
indeed : @xmath102 and @xmath103 by the assumption .
@xmath95 : let @xmath104 be any term of length @xmath31 containing at least one occurrence of each variable @xmath105 and @xmath106 . then @xmath6 is an absorbing term . that leaves only the implication @xmath107 .
we also note that , since we have just shown that the previous two implications always hold , in the later theorems we prove only the implication @xmath107 .
@xmath107 : let @xmath5 , and let @xmath33 be an absorbing term .
let @xmath31 be the length of @xmath6 . then @xmath24 for each @xmath8 , and furthermore , @xmath108 whenever @xmath65 . by the commutativity of @xmath12
, we may write @xmath33 in the form @xmath109 , where @xmath110 .
let us show that @xmath87 for each @xmath8 , @xmath22 .
let @xmath8 , @xmath22 .
let @xmath111 where @xmath78 is chosen so that @xmath112 the length of @xmath113 equals @xmath114 . since @xmath115 and @xmath116 , we get @xmath117 , that is , @xmath113 is well - defined .
further , since @xmath6 is an absorbing term , we have @xmath118 , @xmath119 , @xmath51 , @xmath120 , which gives @xmath121 . finally , note that @xmath122 which proves that @xmath87 .
the proof of @xmath86 is analogous .
conjecture [ conj ] holds when @xmath3 .
@xmath107 : let @xmath5 , and let @xmath33 be an absorbing term .
let @xmath31 be the length of @xmath6 . by the idempotence of @xmath6 it follows that @xmath24 and in fact that @xmath108 whenever @xmath8 and @xmath123 and @xmath124 are positive integers such that @xmath65 .
let @xmath125 .
all we have to prove is that @xmath126 and @xmath127 for each @xmath22 . in the first place , we shall prove that @xmath128 and @xmath129 for each @xmath22 , which shall be needed later . aiming for a contradiction ,
suppose first that @xmath130 for some @xmath22 .
let @xmath36 be the leftmost variable in @xmath33 .
putting @xmath131 at the @xmath50 coordinate and @xmath132 at all the other ones gives @xmath133 let us show that each occurrence of @xmath134 in the above expression can be replaced by @xmath135 without affecting the value of @xmath136 . and indeed , we have @xmath137 which proves the claim .
by iterating this process we ultimately get @xmath138 however , since @xmath6 is an absorbing term , @xmath139 should hold , a contradiction .
this proves that @xmath140 . in an analogous way we obtain that @xmath141 .
again aiming for a contradiciton , suppose now that @xmath142 for some @xmath22 .
it now follows that @xmath143 since @xmath22 and @xmath129 , it is impossible that @xmath144 , since then the value at the right - hand side would belong to @xmath90 , while the value at the left - hand side is @xmath131 . in other words , @xmath145 now , let @xmath146 where @xmath78 is chosen so that @xmath147 . in exactly the same way as in the proof of theorem [ t1 ] , we see that @xmath113 is well - defined and that @xmath148 .
furthermore , we note that each occurrence of @xmath134 in @xmath149 can be replaced by @xmath150 without affecting the value of @xmath149 ; indeed : @xmath151 this enables us to further mimic the proof of theorem [ t1 ] ( in particular , the lines ( [ zaposle ] ) ) , thus obtaining @xmath152 and hence @xmath153
. however , this is exactly the opposite of the supposition ( [ aaa1 ] ) .
this condradiction proves @xmath153 .
the proof of @xmath154 is analogous .
[ tern ] conjecture [ conj ] holds when @xmath2 is an idempotent ternary semigroup .
@xmath107 : let @xmath5 , and let @xmath33 be an absorbing term , where the variables are named in such a way that the leftmost variable is @xmath34 .
let @xmath31 be the length of @xmath6 .
we need to prove that @xmath155 and @xmath156 for any @xmath8 , @xmath22 .
the proof proceeds in nine steps : 1 .
we show that whenever @xmath157 , where @xmath158 is any word of an odd length and @xmath159 of an even length , then @xmath160 .
analogously , we also obtain that , whenever @xmath161 , then @xmath162 .
we show that whenever @xmath163 , where @xmath158 is any word of an even length and @xmath22 , then @xmath164 , and vice versa .
we show that @xmath165 and @xmath166 for any @xmath8 , @xmath22 .
we show that @xmath167 , @xmath168 and @xmath169 for any @xmath8 , @xmath22 .
we show that , whenever @xmath170 is a term such that @xmath171 for all @xmath8 , @xmath22 , then @xmath172 , where @xmath78 is the absolute value of the difference of the number of occurrences of the letter @xmath173 at the odd , respectively even positions in the word @xmath174 .
we show that , whenever @xmath172 for an integer @xmath175 and some @xmath8 , @xmath22 , then @xmath176 .
we show that there exists a positive integer @xmath78 such that @xmath172 and @xmath177 for any @xmath8 , @xmath22 .
we show that @xmath178 for any @xmath8 , @xmath22 .
we show that @xmath155 and @xmath156 for any @xmath8 , @xmath22 .
we now prove these steps . 1 .
since @xmath157 , we obtain @xmath179 .
note that @xmath180 therefore , @xmath181 which gives @xmath160 .
the proof that @xmath161 implies @xmath162 is analogous .
2 . if @xmath163 , then @xmath182 , and now @xmath164 by the previous step .
the other direction is analogous .
3 . let @xmath8 , @xmath22 be given
. denote @xmath183 we have @xmath184 .
note that @xmath185 is a word that starts with @xmath186 , ends with @xmath132 and has no occurrences of two letters @xmath186 next to each other ; furthermore , since @xmath187 , this word reduces to a word that has either @xmath132 or @xmath188 between each two successive occurrences of @xmath186 .
we note @xmath189 that is , each occurrence of @xmath190 in the considered word reduces to @xmath186 .
it follows that @xmath185 reduces to a word @xmath191 , that is , either to @xmath192 or to @xmath193 for some odd positive integer @xmath78 ( @xmath78 has to be odd for these products to be defined ) .
these words further reduce to @xmath194 and @xmath195 , respectively . since @xmath196 and @xmath197 we obtain @xmath165 or @xmath198 . in the first case
, this is what was to be proved . in the second case , since @xmath199 , by step 1 we again obtain @xmath165 .
the proof that @xmath166 is analogous ( or , alternatively , follows from @xmath165 and step 2 ) .
4 . let @xmath8 , @xmath22 be given . by the previous step , we have @xmath200 in an analogous way
, we obtain @xmath201 from these two conclusion we get @xmath202 now , since @xmath203 , step 1 gives @xmath167 .
the proof that @xmath168 is analogous ( or , alternatively , follows from @xmath167 and step 2 ) .
finally , @xmath204 .
let @xmath8 , @xmath22 be given .
we may assume that the leftmost variable in @xmath170 is @xmath105 : indeed , if @xmath170 begins with @xmath106 and @xmath171 for all @xmath205 , then because of step 2 the same holds for the term obtained from the term @xmath206 by moving the leftmost @xmath106 to the end , and this can be repeated until we reach a term that begins with @xmath105 .
we have @xmath207 .
the word @xmath208 is a word that starts with @xmath173 , ends with @xmath132 , and has no occurrences of two letters @xmath173 next to each other .
since @xmath187 , this word reduces to a word that has either @xmath132 or @xmath188 between each two successive occurrences of @xmath173 .
we can write the obtained word in the form @xmath209 with either @xmath210 or @xmath211 at the end , for some positive integers @xmath212 .
+ given a word consisting only of the letters @xmath173 and @xmath132 , by its _ difference _ we shall mean the absolute value of the difference of the number of occurrences of the letter @xmath173 at the odd , respectively even positions in the considered word .
we can assume that the difference of the word @xmath174 is nonzero , since otherwise the conclusion we have to reach is @xmath22 , which holds trivially .
+ notice that the difference of the word @xmath174 equals the difference of the word @xmath208 ( the latter word is obtained by inserting @xmath188 after each occurrence of @xmath173 in the former word , which keeps the parities of the positions at which @xmath173 appears in the former word ) , which in turn equals the difference of the word ( [ sredj ] ) ( replacing @xmath213 by @xmath132 also keeps the considered parities ) , which is evaluated to @xmath214 , then the word ( [ sredj ] ) is @xmath215 ( it can not be @xmath216 because this word has an even length ) and its difference is @xmath123 . now from @xmath217 and step 2 we obtain @xmath218 , which was to be proved . + let @xmath219
. then the word ( [ sredj ] ) is @xmath220 ( there can not be @xmath221 at the end , because then the length would be even ) .
the difference of this word , which is assumed to be nonzero , equals @xmath222 ; therefore , @xmath223 , and in particular , @xmath224 .
now , since @xmath225 and @xmath226 , step 4 gives @xmath227 further , we have @xmath228 therefore , if @xmath225 , then @xmath229 .
further , note that the difference of the latter word equals @xmath230 , that is , this transformation preserves the difference . repeatedly applying
this procedure ultimately leads to @xmath231 or @xmath232 ( depending on whether @xmath233 or @xmath234 ) . in the former case ,
the claim is proved . in the latter case ,
step 2 gives @xmath235 , which again proves the claim .
+ finally , let @xmath236 .
we assume @xmath237 .
let us explain why we are allowed to make this assumption .
if it were @xmath238 , then from @xmath239 ( the case with @xmath240 at the end is similar ) we get , by step 2 , @xmath241 , and now since @xmath238 , everything that follows could be applied as if , informally speaking , it were read from right to left .
+ we claim that there exists a word that belongs to @xmath90 , starts with @xmath173 , ends with @xmath132 , has no occurrences of two letters @xmath173 next to each other , has the same difference as the word ( [ sredj ] ) and is shorter than ( [ sredj ] ) .
+ we first consider the case when @xmath242 or @xmath243 . since @xmath237
, it follows that @xmath244 or @xmath245 , respectively .
we conclude that the minimal possible value of @xmath246 , @xmath247 , is achieved for at least one @xmath98 such that @xmath248 . for such @xmath98 we have @xmath249
this gives the shorter word we were looking for .
indeed , it is enough to check that the obtained word has the same difference as the word ( [ sredj ] ) ( all the other requirements are immediately clear ) .
and indeed , the difference of the obtained word equals @xmath250 which is easily seen to be equal to ( [ diff ] ) . + let now @xmath251 and @xmath252 .
since the word ( [ sredj ] ) is in @xmath90 and @xmath253 , step 4 gives @xmath254 further , we have @xmath255 now , since @xmath256 , by step 2 it follows that the word obtained from this word by moving the letter @xmath132 from the beginning to the end ( and possibly applying @xmath187 at the end , in case that @xmath213 appears there ) also belongs to @xmath90 .
this gives the shorter word we were looking for .
indeed , it is again enough to check only that the obtained word has the same difference as the word ( [ sredj ] ) , which follows by noting that the difference of the obtained word equals @xmath257 + to conclude the proof , we note that repeatedly applying this procedure of shortening the word " ultimately leads to a word of the form treated in one of the cases @xmath258 or @xmath219 , from where we reach the desired conclusion in the already demonstrated way .
let @xmath172 . by step 2 , @xmath259 .
since @xmath167 ( because of step 4 ) , it follows that @xmath260 denote @xmath261 and @xmath262 . since @xmath263 , by step 4 we get @xmath264 .
note that @xmath265 this completes the proof .
let @xmath266 , respectively @xmath267 , denote the number of occurrences of the letter @xmath36 at the odd , respectively even , positions in the word @xmath33 ( we recall that @xmath6 is an absorbing term ) .
then the difference of the word @xmath268 ( @xmath173 is at the @xmath50 coordinate ) equals @xmath269 .
we claim that these differences , for @xmath270 , are coprime ( not necessarily pairwise coprime ) .
suppose the opposite : there exists a prime number @xmath271 such that @xmath272 for each @xmath98 , @xmath270 .
then @xmath273 ( we recall that @xmath31 is the length of @xmath6 ) , which is a contradiction
. therefore , the considered differences are coprime .
+ by step 5 , if @xmath246 is any of these differences , then @xmath274 for all @xmath205 .
if @xmath246 and @xmath275 are any two of these differences , then @xmath276 therefore , @xmath172 whenever @xmath78 is any linear combination with nonnegative integer coefficients of the considered differences .
since these differences are coprime , any large enough positive integer can be represented as a linear combination of them . in particular
, there indeed exists a positive integer @xmath78 such that @xmath172 and @xmath177 .
let @xmath277 be the least positive integer such that @xmath278 and @xmath279 for any @xmath8 , @xmath22 ( such a number exists by the previous step ) .
we need to prove that @xmath280 . aiming for a contradiction ,
suppose that @xmath281 .
we shall prove that @xmath282 for any @xmath8 , @xmath22 , which , together with the assumed @xmath278 , contradicts the minimality of @xmath277 .
+ we first treat the case @xmath283 .
let @xmath8 , @xmath22 be given , and let us prove that @xmath178 .
we have @xmath284 and @xmath285 . by step 6 , we have @xmath286 and @xmath287 . since , by step 4 , @xmath167 and @xmath168 , we conclude @xmath288 further , we have @xmath289 this proves the case @xmath283
. + let now @xmath290 .
let @xmath8 , @xmath22 be given , and let us prove that @xmath282 .
denote @xmath291 and @xmath292 . by step 3
, we have @xmath263 .
therefore , @xmath293 and @xmath294 . by step 6 , we have @xmath295 and @xmath296 . by step 4
, we have @xmath297 .
therefore , @xmath298 further , we have @xmath299 ( between the last and the next to last row we used the fact that after each @xmath300 there is another @xmath301 following , and thus , because of @xmath302 , we may simply erase each such @xmath300 . )
this completes the proof .
let @xmath8 , @xmath22 be given . by the previous step , we have @xmath178 . by step 2 ,
we now obtain @xmath155 and @xmath156 , which was to be proved . the proof of theorem [ tern ] is thus finished . for the end
, we prove a proposition that shows that the requirement that @xmath2 is idempotent from the previous theorem is , in a way , not so restrictive as it might seem to be .
assume that conjecture [ conj ] holds for all idempotent @xmath0-ary semigroups .
then conjecture [ conj ] holds in general .
as before , it is enough to prove only the direction @xmath107 .
let @xmath5 , and let @xmath33 be an absorbing term .
let @xmath31 be the length of @xmath6 . then the algebra @xmath303 , where @xmath304 , is a @xmath31-ary idempotent semigroup .
the term @xmath6 is also an absorbing term for @xmath1 in @xmath305 , that is , @xmath306 .
therefore , by the assumed special case of conjecture [ conj ] , @xmath307 and @xmath308 for each @xmath8 , @xmath22 .
of course , the same also holds in @xmath2 . from here
it is easy to prove that @xmath87 and @xmath86 for each @xmath8 , @xmath22 ; indeed : @xmath309 and an analogous reasoning shows @xmath86 .
the author would like to thank the two anonymous referees for thorough reading of the paper and many useful comments , and in particular for a suggestion about a possible application of the results from the paper mentioned in the introduction .
the research was supported by the ministry of science and technological development of serbia ( project 174006 ) and by the provincial secretariat for science and technological development , autonomous province of vojvodina ( project ordered structures and applications " ) .
10 l. barto , the dichotomy for conservative constraint satisfaction problems revisited , in : _ proceedings of the 26th annual ieee symposium on logic in computer science ( lics 2011 ) _ , toronto , ontario , canada , 2011 , pp . 301310 .
l. barto & a. kazda & j. buln , the distance from congruence distributivity to near unanimity , _ general algebra and its applications ( gaia2013 ) _ , melbourne , australia , 2013 .
l. barto & m. kozik , congruence distributivity implies bounded width , _ siam journal on computing _ * 39 * ( 2010 ) , 15311542
. l. barto & m. kozik , constraint satisfaction problems of bounded width , in : _ proceedings of the 50th symposium on foundations of computer science ( focs09 ) _ , 2009 , pp . 595603 .
l. barto & m. kozik , absorbing subalgebras , cyclic terms , and the constraint satisfaction problem , _ logical methods in computer science _
* 8 * ( 2012 ) , 126
. l. barto & m. kozik & t. niven , graphs , polymorphisms and the complexity of homomorphism problems , in : _ proceedings of the 40th annual acm symposium on theory of computing ( stoc08 ) _ , new york , usa , 2008 , pp
. 789796 . l. barto & m. kozik & t. niven , the csp dichotomy holds for digraphs with no sources and no sinks ( a positive answer to a conjecture of bang - jensen and hell ) , _ siam j. comput . _
* 38 * ( 2009 ) , 17821802 .
a. a. bulatov , complexity of conservative constraint satisfaction problems , _ acm trans .
* 12 * ( 2011 ) , art .
24 , 66 pp .
a. bulatov & p. jeavons & a. krokhin , classifying the complexity of constraints using finite algebras , _
siam j. comput . _
* 34 * ( 2005 ) , 720742 .
j. buln , decidability of absorption in relational structures of bounded width , _ algebra universalis _ * 72 * ( 2014 ) , 1528
. b. a. davey & l. heindorf & r. mckenzie , near unanimity : an obstacle to general duality theory , _ algebra universalis _ * 33 * ( 1995 ) , 428439 . t. feder & m. y. vardi , the computational structure of monotone monadic snp and constraint satisfaction : a study through datalog and group theory , _ siam j. comput .
_ * 28 * ( 1999 ) , 57104 .
p. jeavons & d. cohen & m. gyssens , closure properties of constraints , _
j. acm _ * 44 * ( 1997 ) , 527548 .
a. kazda , how to decide absorption , _
the 4th novi sad algebraic conference ( nsac2013 ) _ , novi sad , serbia , 2013 .
m. marti , the existence of a near - unanimity term in a finite algebra is decidable , _ j. symbolic logic _ * 74 * ( 2009 ) , 10011014 .
m. marti & r. mckenzie , existence theorems for weakly symmetric operations , _ algebra universalis _ * 59 * ( 2008 ) , 463489 . | the notion of absorption was developed a few years ago by barto and kozik and immediately found many applications , particularly in topics related to the constraint satisfaction problem .
we investigate the behavior of absorption in semigroups and @xmath0-ary semigroups ( that is , algebras with one @xmath0-ary associative operation ) . in the case of semigroups
, we give a simple necessary and sufficient condition for a semigroup to be absorbed by its subsemigroup .
we then proceed to @xmath0-ary semigroups , where we conjecture an analogue of this necessary and sufficient condition , and prove that the conjectured condition is indeed necessary and sufficient for @xmath1 to absorb @xmath2 ( where @xmath2 is an @xmath0-ary semigroup and @xmath1 is its @xmath0-ary subsemigroup ) in the following three cases : when @xmath2 is commutative , when @xmath3 and when @xmath2 is an idempotent ternary semigroup . |
cooperation is vital for the maintenance of public goods in human societies @xcite . but according to darwin s theory of evolution , competition rather than cooperation ought to drive our actions .
the reconciliation of this theory with the fact that cooperation is widespread in human societies , as well as with the fact that it is much more common in nature as one might expect , is one of the most persistent challenges in evolutionary biology and social sciences @xcite .
past decades have seen the paradigm of punishment rise as one of the more successful strategies by means of which cooperation might be promoted @xcite .
indeed , punishment is also the principle tool of institutions in human societies for maintaining cooperation and otherwise orderly behavior @xcite .
however , punishment is costly , and as such it reduces the payoffs of both the defectors as well as of those that exercise the punishment , hence yielding an overall lower income and acting as a drain on social welfare .
thus , understanding the emergence of costly punishment is crucial for the evolution of cooperation @xcite . while recent research confirms that punishment is often motivated by negative personal emotions such as anger or disgust @xcite , raihani and mcauliffe
have shown also that the decision to punish is often motivated with the aversion of inequity in mind , rather than by the desire for reciprocity @xcite .
although prosocial punishment is widespread in nature @xcite , it is unlikely that cooperators are willing to commit permanently to punishing wrongdoers .
for that , the action is simply to costly , and hence some form of abstinence is likely , also to avoid unwanted retaliation .
several research groups have recently investigated these and related up and down sides of punishment @xcite .
for example , it was shown that cooperators punish defectors selectively depending on their current personal emotions , even if the number of defectors is large @xcite . more often than not , however , whether or not to punish depends on the whiff of the moment and is thus a fairly random event .
motivated by these observations , we have recently shown that sharing the effort of punishment in a probabilistic manner can significantly lower the vulnerability of costly punishment and in fact help stabilize costly altruistic strategies @xcite . here
we drop the assumption that cooperators who do punish defectors do so uniformly at random .
instead , we account for the diversity in punishment , taking into account the fact that some individuals are more likely to punish , while others punish only rarely .
more specifically , we introduce different threshold levels for punishment , which ultimately introduces different classes of cooperators that punish defectors .
the assumption of diverse players is not just a realistic hypothesis , but in general it is firmly established that it also has a decisive impact on the evolution of public cooperation @xcite .
motivated by this fact , we therefore study a spatial public goods game with defectors and different types of punishing cooperators .
while previously we have demonstrated the importance of randomly shared punishment @xcite , we here approach a more realistic scenario by assuming that each type of cooperators will punish with a different probability .
our goal is to determine whether a specific class of punishing cooperators will be favored by natural selection , or whether despite the competition among them synergistic effects will emerge .
as we will show , the evolution is governed by a counterintuitive selection mechanism , depending further on the synergistic effects of cooperative behavior . however , before presenting the main results in detail , we proceed by a more accurate description of the studied spatial public goods game with different punishing strategies .
we consider a population of individuals who play the public goods game on a square lattice of size @xmath0 with periodic boundary conditions .
we assume that the game is contested between @xmath1 classes of cooperators ( @xmath2 , @xmath3 , @xmath4 , @xmath5 ) and defectors ( @xmath6 ) .
independently of the class a cooperator belongs to , it contributes an amount @xmath7 to the common pool , while defectors contribute nothing .
after the sum of all contributions in the group is multiplied by the enhancement factor @xmath8 , the resulting amount is shared equally among all group members .
moreover , cooperators with strategy @xmath9 ( @xmath10 ) choose to punish defectors with a probability @xmath11 if the latter are present .
as a result , each defector in the group is punished with a fine @xmath12 , while all the cooperators who participated in the punishment equally shared the associated costs .
in particular , each punishing cooperator bears the cost @xmath13 , where @xmath14 and @xmath15 are the number of cooperators and punishers in the group , respectively . we emphasize that a cooperator who decides to punish bears the same cost independently of the class it belongs to .
thus , here the strategy @xmath16 only determines how frequently a cooperator is willing to punish defectors .
nevertheless , it is worth pointing out that @xmath2 never punish and thus correspond to traditional second - order free - riders because they enjoy the benefits of punishment without contributing to it @xcite . on the other extreme ,
cooperators belonging to the @xmath5 class punish always when defectors are present in the group . since each player on site
@xmath17 with von neumann neighborhood is a member of five overlapping groups of size @xmath18 , in each generation it participates in five public goods games and obtains its total payoff @xmath19 , where @xmath20 is the payoff gained from group @xmath21 .
subsequently , a player @xmath17 , having strategy @xmath22 , adopts the strategy @xmath23 of a randomly chosen neighbor @xmath24 with the probability @xmath25},\ ] ] where @xmath26 denotes the amplitude of noise @xcite . without loosing generality and to ensure continuity of this line of research @xcite we set @xmath27 , meaning that it is very likely that the better performing players will pass their strategy to their neighbors , yet it is also possible that players will occasionally learn from a less successful neighbor . to conclude the description of this public good game
, we would like to emphasize that different @xmath9 classes represent different strategies , as our goal is to explore how the willingness to punish evolves at specific parameter values .
the model is studied by means of monte carlo simulations .
initially , defectors randomly occupy half of the square lattice , and each type of cooperators randomly @xmath28 of the rest of the lattice . during one full monte carlo step ( mcs ) , all individuals in the population receive a chance once on average to adopt another strategy . depending on the proximity to phase transition points and the typical size of emerging spatial patterns ,
the linear system size was varied from @xmath29 to @xmath30 and the relaxation time was varied from @xmath31 to @xmath32 mcs to ensure proper statistical accuracy .
the reported fractions of competing strategies were determined in the stationary state when their average values became time - independent .
alternatively , we have averaged the outcomes over @xmath33 independent runs when the system terminated into a uniform absorbing state .
and the probability to punish @xmath34 , as obtained for a low multiplication factor @xmath35 in the original model proposed in @xcite , where a uniform probability to punish was assumed for all cooperators .
note that both @xmath12 and @xmath34 have a non - monotonous impact on the fraction of cooperators.,width=264 ] for the sake of comparison , we first present the fraction of cooperators in dependence on the punishment fine @xmath12 and the probability to punish @xmath34 at a low @xmath36 value , as obtained in the original probabilistic punishment model , where cooperators punish uniformly at random @xcite .
figure [ fig1 ] illustrates that the fraction of cooperators first increases , reaches its maximum , but then again decreases , as the values of @xmath12 and @xmath34 increase along the diagonal on the @xmath37 plane .
increasing one of these parameters , while the other is kept constant , returns to the same observation .
both @xmath12 and @xmath34 thus have a non - monotonous impact on the fraction of cooperators , which is closely related with the fact that @xmath12 characterizes not only the level of punishment but also its cost .
accordingly , too high values of @xmath12 involve too high costs stemming from the act of punishing .
it is worth pointing out that @xmath35 , which is used in fig .
[ fig1 ] , is a relatively low value of the multiplication factor at which the non - monotonous dependence can still be observed . in comparison with the results obtained for larger values of @xmath36 as used in ref .
@xcite , however , the current plot features a significantly narrower @xmath34 region where full cooperation is possible when @xmath12 is sufficiently large . similarly , there is a limited region of intermediate @xmath12 values where cooperators that punish severely can beat defectors . based on these observations , in the present model
we thus explore if there is an evolutionary selection among different punishing strategies as they compete against the defectors simultaneously , or if there is indeed cooperation in the common goal to deter defectors . when they start fighting with defectors simultaneously .
panel ( c ) shows an enlarged part of panel ( a ) at low @xmath12 values , when cooperation becomes dominant over defection . to present the overall level of cooperation in the population
, the cumulative fraction of @xmath9 strategies is also shown ( denoted by @xmath38 ) . for comparison , in panel ( b ) we have also plotted the resulting fraction of cooperator classes when they fight against defectors individually . as in panel ( c ) , panel ( d ) shows an enlarged part of panel ( b ) at a specific interval of @xmath12 .
the multiplication factor in all panels is @xmath35.,width=321 ] for an intuitive overview , we set @xmath39 and investigate how the six types of punishing strategies compete and potentially cooperate with each other in the presence of defectors .
the general conclusion , however , is robust and remains valid if we use other values of @xmath1 .
using the same @xmath35 as in fig .
[ fig1 ] , the panels of fig .
[ fig2 ] summarize our main findings .
the first panel shows the fractions of strategies in the final state in dependence of the punishment fine @xmath12 when different punishing strategies fight against defectors simultaneously .
for clarity , we have also plotted the accumulated fraction of punishing strategies .
in contrast to the uniform punishing model , we can see that the total fraction of cooperators should increase monotonously with increasing @xmath12 . as fig .
[ fig2](a ) illustrates , cooperators can survive when @xmath40 , and become dominant over @xmath41 ( see also the enlarged part in fig .
[ fig2](c ) ) .
we should stress , however , that not all types of cooperators can survive at equilibrium , even if cooperators take over the whole population .
it turned out that there are some `` weak '' classes of cooperators who go extinct before defectors die out , while other classes of cooperators survive . for a more in - depth explanation
, the vitality of punishing classes can be estimated if we let them fight against defectors individually .
the outcomes of this scenario are summarized in fig .
[ fig2](b ) .
results presented in this panel suggest that there are punishing classes who can dominate for all high @xmath12 values , while others become vulnerable as we increase @xmath12 .
more interestingly , however , there are mildly punishing strategies who can survive only due to the support of the more successful punishing strategies .
for example , for @xmath42 classes @xmath43 and @xmath44 can outperform defectors , while @xmath45 disappear when they fight against defectors individually [ fig .
[ fig2](b ) and ( c ) ] .
but when all punishing strategies are on the stage then @xmath45 players can survive as well .
this effect is more spectacular for the second - order free riding @xmath2 class , who would die out immediately at such a low synergy factor @xmath36 if they face defectors alone .
but now , especially at high @xmath12 values , their ratio becomes considerable .
this indicates that some less viable classes of cooperators can survive because of the support of more viable punishing strategies via an evolutionary selection mechanism which has a biased impact on the evolution of otherwise competing strategies . to demonstrate the underlying mechanism behind the above observations , we present a series of snapshots of strategy evolutions starting from different prepared initial states
the comparative analysis is plotted in fig .
[ fig3 ] , where all runs were obtained for @xmath42 and @xmath35 . in the first row
, we demonstrate how the class of @xmath43 punishing strategy can prevail over defectors .
initially , only a tiny portion of @xmath43 cooperators is launched in the sea of defectors [ the fraction of @xmath43 is @xmath46 , see panel ( a ) ] .
still , @xmath43 cooperators can expand gradually and invade the whole available territory [ shown in panels from ( a2 ) to ( a4 ) ] .
the second row , which was taken at the same parameter values , demonstrates clearly the vulnerability of the @xmath45 class against defectors . despite of the fact that they occupy the majority of the available room at the beginning , shown in panel ( b1 ) ,
still , they will be gradually crowded out by defector players .
the final state , shown in panel ( b4 ) , highlights that such a rare punishment activity represented by @xmath45 class is ineffective against defectors at the applied synergy factor @xmath36 .
the third row , where all previously mentioned strategies are present at the beginning , illustrates a completely different scenario . here
we start from a balanced initial state where half of the lattice sites is occupied by @xmath45 and @xmath43 strategies , while the other half is filled by defectors . as panels ( c1 ) to ( c4 )
illustrate , defectors will gradually go extinct while `` weak '' @xmath45 cooperators survive and occupy almost half of the available territory in the final state .
we note that there is a neutral drift between punishing strategies in the absence of defectors , which will result in a homogeneous state where the probability to arrive to one of the possible final destinations is proportional to the initial portion of a specific class at the time defectors die out @xcite .
this evolutionary outcome indicates that although @xmath45 players are , as an isolated strategy , weak against defector players , they can nevertheless survive because of the assistance of the strong @xmath43 strategy even if the initial fraction of the later is modest . in the fourth row , however , when we arrange a similar setup but replaced weak @xmath45 players with also weak @xmath47 players , the final state will always be the full @xmath43 state . here
, the presence of strong @xmath43 players does not yield a relevant support to @xmath47 players who therefore die out , and subsequently the system returns to the scenario illustrated in panels ( a1 ) to ( a4 ) . and @xmath35 .
the first row shows the case when just a few @xmath43 cooperators are initially present among defectors .
it can be observed that even under such unfavorable initial conditions the @xmath43 strategy can successfully outperform defectors .
the second row feature a similar experiment with the @xmath45 strategy , which fails to survive among defectors even though the latter are initially in minority .
the third row illustrates cooperation among strategies @xmath45 and @xmath43 , which together dominate the whole population even though @xmath45 alone would fail under the same conditions ( see second row ) .
we note that a neutral drift starts when defectors die out , as explained in the main text .
the fourth row demonstrates , however , that the cooperation among different punishing strategies illustrated in the third row is rather fragile .
if initially the strategy @xmath45 is replaced by strategy @xmath47 , then the later simply die out and subsequently the whole evolution becomes identical to the one shown in the first row , where strategy @xmath43 alone outperforms all defectors . for clarity , here
the employed system size is small with just @xmath48 players.,width=321 ] the key point , which explains the significantly different trajectories for mildly punishing strategies is based on the difference of invasion velocities between the competing strategies . to demonstrate the importance of invasion velocities
, we monitor how the fraction of strategies evolves in time when we launch the system from a two - strategy state where both strategies form compact domains .
following the previously applied approach illustrated in fig .
[ fig3 ] , we compare the strategy invasions between @xmath49 , @xmath50 , and between @xmath51 strategies .
the comparison of these different cases is plotted in fig .
as expected , both @xmath47 and @xmath45 loose the lonely fight against defectors , while @xmath43 will eventually crowd out defectors .
note that there is only a very slight increase during the early stages of the evolutionary process that can be observed for all cases , independently of the final outcome .
this is because straight initial interfaces can provide a strong temporary phalanx for every punishing strategy .
nevertheless , when this interface becomes irregular due to invasions the individual weakness of @xmath47 and @xmath45 strategies reveals itself . still , there is a significant difference between their trajectories .
namely , strategy @xmath45 is able to resist for a comparatively long time , which gives strategy @xmath43 enough time to crowd out defectors .
on the other hand , strategy @xmath47 is a too easy prey for defectors , which is why they die out faster than the strategy @xmath43 is able to eliminate all defectors . ultimately thus , strategy @xmath45 can benefit from cooperation with strategy @xmath43 , while strategy @xmath47 is unable to do the same . ,
@xmath45 and @xmath43 , against defectors in dependence on time .
note that initially only one cooperative strategy and defectors are present , using the same initial conditions as illustrated in fig .
[ fig3 ] . positive value of @xmath52 indicates the invasion of cooperator strategy while its negative value suggests invasion to the reversed direction .
note that while both @xmath47 and @xmath45 strategies ultimately loose their battle , the latter is able to prevail significantly longer .
this enables an effective help of strategy @xmath43 when they compete against defectors together , as illustrated in panels ( c1 ) to ( c4 ) in fig .
[ fig3].,width=321 ] in the remainder of this work , we focus on the parameter region where cooperators are able to coexist with defectors without applying punishment .
namely , if the synergy factor exceeds @xmath53 , then pure cooperators ( cooperators that do not punish ) can survive permanently alongside defectors due to network reciprocity @xcite . evidently , the presence of punishers can of course still elevate the overall cooperation level and defectors can be effectively crowded out from the population @xcite . here
the main question is thus how the different punishing strategies will share the available space .
the results are summarized in the left panel of fig .
[ r4 ] , as obtained for the representative value of @xmath54 .
it can be observed that , when all the different types of punishing strategies fight against defectors simultaneously , then cooperators can dominate the whole population above a threshold value @xmath55 .
however , to evaluate these final outcomes adequately , we need to know the individual relations between each particular cooperative strategy and defectors on a strategy - versus - strategy basis . therefore , as for the previously presented low @xmath36 case in fig .
[ fig2 ] , in the right panel of fig .
[ r4 ] we also show the stationary fractions of different cooperators classes when they compete against defectors individually .
results presented in panel ( b ) highlight that too large @xmath12 values could be detrimental for the @xmath45 , @xmath44 and the @xmath43 strategy .
this is the so - called `` punish , but not too hard '' effect , where too large costs of sanctioning do more damage to those that execute punishment than the imposed fines do damage to the defectors @xcite . a direct comparison with the results presented in panel ( a ) demonstrates clearly that we can observe a similar cooperation among punishing strategies as we have reported before for the low @xmath36 case , in particular because all the mentioned mildly punishing strategies can survive even at a high @xmath12 value . on the other hand
, a conceptually different mechanism can be observed in the small @xmath12 region , which is reminiscent of what one would actually expect from a selection process .
more specifically , panel ( a ) of fig .
[ r4 ] shows that at @xmath56 only strategy @xmath45 survives and coexists with @xmath6 while all the other punishing strategies die out .
the latter players are those , who could survive individually with defectors but should die out because of the presence of a more effective ( @xmath45 ) strategy .
interestingly , the mentioned selection mechanism can work most efficiently when the leading strategy is less efficient against defectors .
right panel of fig .
[ r4 ] shows that @xmath45 would be unable to crowd out strategy @xmath6 at these @xmath12 values , while a @xmath6-free state could be obtained at higher @xmath12 value . in the latter case ,
when @xmath45 is too powerful , then this strategy beats defectors too fast which allows other punishing strategies to survive : this is similar to what we have observed in the third row of fig .
but when @xmath45 is less effective at smaller @xmath12 values then the presence of surviving @xmath6 players enables @xmath45 players to play out their superior efficiency if comparing to other punishing strategies .
thus , depending on the key parameter values , most prominently the multiplication factor @xmath36 and the punishment fine @xmath12 , the different punishing strategies can either cooperate with each other or compete against each other in the spatial public goods game .
we have introduced and studied multiple types of punishing strategies that sanction defectors with different probabilities .
the fundamental question that we have addressed is whether there exists a selection mechanism which would result in an unambiguous victor when these strategies compete against defectors .
we have shown that the answer to this question depends sensitively on the external conditions , in particular on the value of the multiplication parameter @xmath36 .
if the public goods game is demanding due to a low value of @xmath36 , then the pure payoff - driven individual selection provides a helping hand to those punishing strategies that would be unable to survive in an individual competition against defectors .
in particular , we have demonstrated that the failure or success of a specific punishing strategy could depend sensitively on the relation of invasion velocities between specific punishing strategies and the defectors .
accordingly , if the loosing punishing strategy can delay the complete victory of defectors sufficiently long , then a more successful punishing strategy has a chance to wipe out defectors first .
this is an example of the cooperation between different punishing strategies . on the other hand , in a less demanding environment , characterized by a higher multiplication factor
, a different kind of relation can emerge .
while the previously summarized cooperation between punishing strategies is still possible , there also exist parameter regions where competition is the dominant mode , and indeed there is always a single and unambiguous victor among the different classes of punishers .
interestingly , we have shown that this happens when the fittest punishing strategy is not effective enough to beat defectors completely . instead , by carefully taming the defectors , they help to reveal the advantages of other punishing strategies .
as we have shown , the key point here is again the relation between the invasion velocities .
namely , a too intensive invasion will decimate defectors too fast and the advantage of specific punishing classes will remain forever hidden . therefore , in contrast to intuitive expectation , the social diversity of cooperators in terms of their relations with defectors could be the result of an effective selection mechanism .
we hope that this research will contribute relevantly to our understanding of the emergence of diversity among competing strategies , as well to their role in determining the ultimate fate of the population .
this work was supported by the fundamental research funds of the central universities of china , the hungarian national research fund ( grant k-101490 ) , the slovenian research agency ( grant p5 - 0027 ) , and by the deanship of scientific research , king abdulaziz university ( grant 76 - 130 - 35-hici ) . | inspired by the fact that people have diverse propensities to punish wrongdoers , we study a spatial public goods game with defectors and different types of punishing cooperators . during the game , cooperators punish defectors with class - specific probabilities and subsequently share the associated costs of sanctioning .
we show that in the presence of different punishing cooperators the highest level of public cooperation is always attainable through a selection mechanism .
interestingly , the selection not necessarily favors the evolution of punishers who would be able to prevail on their own against the defectors , nor does it always hinder the evolution of punishers who would be unable to prevail on their own .
instead , the evolutionary success of punishing strategies depends sensitively on their invasion velocities , which in turn reveals fascinating examples of both competition and cooperation among them .
furthermore , we show that under favorable conditions , when punishment is not strictly necessary for the maintenance of public cooperation , the less aggressive , mild form of sanctioning is the sole victor of selection process .
our work reveals that natural strategy selection can not only promote , but sometimes also hinder competition among prosocial strategies . |
_ cgro_-egret established the predominance of flat spectrum radio quasars ( fsrq ) as extragalactic sources of high - energy @xmath0-rays @xcite .
these are believed to be powered by supermassive black holes ejecting large amounts of material along relativistic jets perpendicular to accretion disks and producing high energy particles in these jets .
observations at the highest photon energies shown a predominance of bl lac objects over fsrq at @xmath1 .
although these observations may be biased by the horizon limitation due to pair absorption with extragalactic background light ( ebl ) , the _ fermi _
@xmath0-ray space telescope , more sensitive and more responsive to gev photons than egret , reported a larger fraction of bl lac objects among blazars in its bright source list release ( 0fgl ) @xcite .
the wider spectral coverage of _ fermi _ is allowing detailed studies of the intrinsic properties of various types of active galactic nuclei ( agn ) , cosmic evolution and their use as ebl tracers @xcite .
the extrapolation of fsrq to high radio frequencies makes the coincidence of foreground _ wmap _ sources with @xmath0-ray blazars expectable .
furthermore , it is clear than most of the 390 _ wmap _ sources are agns @xcite .
these emitters are dominated in the mm - wave band by the same synchrotron component observed at lower frequencies and can be expected to emit photons of very high energy , related to the hardest part of the synchrotron emission .
_ wmap _ sources have fluxes above 0.1 jy at frequencies @xmath2 , making them suitable positional references for large millimeter telescopes .
known common _
wmap_/egret sources , found up to @xmath3 , are relatively nearby analogs of more distant blazars detectable in principle by _ fermi _ up to @xmath4 and by the large millimeter telescope ( lmt @xcite ) at @xmath5 , if such an object could have existed at such an early phase of the universe .
the combination of _ fermi _ and lmt data promises the availability of a sample of thousands of objects up to the highest redshifts ideally suited to study radio loud agn evolution and the connection of the ebl with the star formation history of the universe . | we established a sample of millimeter - wave and @xmath0-ray bright active galactic nuclei matching the _ wmap _ catalog with the egret catalog , highest energy photons and the _ fermi _ bright source list .
we have monitored over 80 of these objects in the near infrared , obtaining over 2000 jhks data points directly comparable with _ fermi _ data .
we present examples of correlated near infrared and @xmath0-ray activity of known blazars and recently identified sources . |
most of the effort in the research on irreversible adsorption is focused on flat homogeneous surfaces as it can be further directly exploited in material sciences .
however , experiments using fractal - like collectors were also made @xcite , though preparation of such surfaces is complicated . on the other hand , there are many porous media in nature and adsorption may play an important role there @xcite .
for example , coral fractal - like structure helps them to catch plankton effectively @xcite .
adsorption on fractal collectors might also be applied in environmental protection in designing effective water or air filters @xcite .
hard spheres packing has a rich research history , both in physics and maths .
the greatest interest has been shown in the maximal possible packing @xcite .
the most common property of a given packing is its density . in @xmath3
dimensional euclidean space , it is given by @xmath4 where @xmath5 is a number of spheres in a unit volume , and @xmath6 is a volume of @xmath3 dimensional sphere having radius @xmath7 .
@xmath8 is the euler gamma function .
presented work focuses on the properties of maximal random coverages , built of spherical particles on flat collectors having fractal dimension of a non - integer value .
earlier , similar studies concentrated mostly on adsorption on lattice collectors .
their popularity was powered by a number of analytic results obtained , e.g. @xcite .
other investigations concerning fractal collectors concentrate on diffusion properties @xcite or adsorption on rough surfaces @xcite .
fractal structures formed in the adsorption processes were also observed @xcite .
dimensional properties of adsorption layers were studied mostly for collectors having integer dimensions , with special attention paid to two - dimensional case , e.g. @xcite , due to its potential application in chemistry and material science . here , the only analytically solved case is a one - dimensional problem known also as car parking problem for which @xmath9 @xcite .
the analysis of hard spheres random packing for @xmath10 , nicely reviewed in @xcite , shows that maximal random coverage ratio decreases with the growth of collector s dimension : @xmath11 with @xmath12 and @xmath13 @xcite .
there are also some characteristics , which do not depend qualitatively on dimension .
for example the two - point density correlation function for maximal random coverages is known to have superexponential decay for large distances @xcite @xmath14 and logarithmic singularity when particles are in touch @xcite @xmath15 the aim of this work is to check if the mentioned characteristics are valid for adsorption on collectors having non - integer fractal dimensions .
the maximal coverages has been obtained here numerically by using random sequential adsorption ( rsa ) algorithm @xcite described in detail in section [ sec : model ] .
the kinetics of rsa simulation has also been of interest to the authors of this work , as it affects estimation of @xmath16 . for sufficiently long simulation time the coverage ratio @xmath17 scales with collector s dimension @xcite : @xmath18 here @xmath19 .
importance of the above relation is increased by the fact that it is also valid for other molecules , more complex than simple spheres @xcite . to model collectors having dimension from a range of @xmath20 the sierpinski s triangle and , carpet - like fractals
adsorption for @xmath2 is studied using general cantor set .
maximal random coverages are generated using rsa algorithm , which has been successfully applied to study colloidal systems .
it is based on independent , repeated attempts of adding sphere to a covering film .
the numerical procedure is carried out in the following steps : i : : a virtual sphere is created with its centre position on a collector chosen randomly according to the uniform probability distribution ; ii : : an overlapping test is performed for previously adsorbed nearest neighbours of the virtual particle .
the test checks if surface - to - surface distance between spheres is greater than zero ; iii : : if there is no overlap the virtual particle is irreversibly adsorbed and added to an existing covering layer
. its position does not change during further calculations ; iiii : : if there is an overlap the virtual sphere is removed and abandoned .
attempts are repeated iteratively .
their number is usually expressed in a dimensionless time units : @xmath21 where @xmath22 is a number of attempts and @xmath23 denotes an area of a single sphere projection on a collector .
typically , @xmath24 is a collector area .
although formally the area of any fractal having @xmath25 is zero , @xmath24 denotes here an area of an overlapping square and is equal to @xmath26 , where @xmath27 is a radius of an adsorbed sphere .
collectors are modeled as a successive iteration of a given fractal . for each collector type
at least @xmath28 independent numerical experiments have been performed .
sample coverages are presented in fig.[fig : examples ] .,width=302 ] dimension of a given fractal is estimated using relation @xmath29 where @xmath30 is a number of identical smaller parts created in a single iteration and @xmath31 is a scale change . in case of cantor dust , for example , after one iteration four smaller squares appear after one iteration @xmath32 with a side length three times smaller than in the original square @xmath33 .
therefore , dimension of cantor dust is @xmath34 coverage ratio of a given adsorption layer is determined by using random sampling of collector s points and checking whether they are covered by any , previously adsorbed sphere : @xmath35 where @xmath36 is a number of covered points after a dimensionless time @xmath37 ; @xmath38 denotes total number of sampled points .
here @xmath39 , which provides statistical error at the level of @xmath40 .
there are two factors affecting maximal random coverage ratio obtained during simulation .
first , the collector used is not an actual fractal but only a two - dimensional approximation and the number of adsorbed spheres clearly depends on the level of this approximation ( see fig.[fig : examples ] ) .
second , the maximal random coverage is defined for a collector having infinite size and is achieved after an infinite simulation time . to handle the first factor ,
the successive iterations of fractals have been used .
the number of adsorbed particles becomes stable at about the @xmath41 fractal iteration ( see fig.[fig:2dni ] ) .
similarly , the coverage calculated using eq.([eq : qt ] ) does not change thereafter , too ( fig.[fig:2dti ] ) . therefore , using the @xmath42 iteration has been assumed to be save and to give reliable results . as a function of fractal s iteration .
iteration zero is a solid triangle in case of the sierpinski s triangle and a square in all the other cases.,width=302 ] to obtain maximal random coverage ratio after an infinite time , the rsa kinetics model has to be used .
the analysis of numerical experiments shows that rsa for spheres in integer dimensions obeys the feder s law ( [ eq : feder ] ) .
the results presented in fig.[fig : federlaw ] confirm that for almost all the cases analysed in this study .
additionally , the difference between observed maximal number of adsorbed particles for @xmath43 and the asymptotic value does not exceed @xmath44 .
coverage does not grow asymptotically only in case of adsorption on the sole incoherent collector studied , the cantor dust . in this case , the maximal possible number of @xmath45 particles is adsorbed after relatively short time .
this is due to geometrical relations between collector size and spheres diameter ( see fig.[fig : cantordust ] ) .
after four iterations , each of the remaining squares has a side length of @xmath46 and can adsorb only one sphere .
moreover , due to the large distance between parts of the fractal , particle adsorbed on one square can not block absorption events on its neighbouring squares .
therefore , each square contains a single sphere .
it is noteworthy that further iterations of the fractal do not affect the above reasoning .
therefore , after catching @xmath47 spheres no further adsorption acts are possible .
investigating the relationship between maximal random coverage ratio and fractal dimension is the main purpose of this study .
although , there are number of studies providing lower and upper limits for a saturated coverage , e.g. @xcite , the best direct fit so far is given by eq.([eq : fit1 ] ) .
it perfectly matches simulation data for @xmath10 @xcite .
however , it does not work well for @xmath48 where @xmath49 and @xmath50 where @xmath51 . to obtain better approximation
, we added next order term in the nominator of relation ( [ eq : fit1 ] ) : @xmath52 figure [ fig : qd16 ] presents plot of @xmath16 where values for @xmath1 were obtained with ( [ eq : qt ] ) and ( [ eq : feder ] ) whereas values for @xmath53 were taken from @xcite . )
fit ( @xmath54 and @xmath55 ) , whereas dashed line is the ( [ eq : fit2 ] ) fit with @xmath56 , @xmath57 and @xmath58.,width=302 ] the fit defined by eq.([eq : fit2 ] ) matches the data for @xmath0 and is characterized by correlation coefficient @xmath59 .
furthermore , the parameter @xmath60 corresponds very well to random saturation density for @xmath50 .
the correction given by the third term @xmath61 = @xmath62 is important mostly for large @xmath3 values .
therefore , simulations for higher dimensions should determine whether @xmath63 should remain or not . two - point correlation functions for fractal collectors are presented in fig.[fig : cor ] . again ,
except for the case of cantor dust collector , all the plots look typical ; for @xmath64 , however , the specific fractal structure is manifested by a number of local extremes .
therefore , the pure superexponential decay ( [ eq : corfast ] ) can not be observed . for small values of @xmath27 ,
the logarithmic singularity ( [ eq : corlog ] ) appears just as for adsorption in higher dimensions . in case of the cantor dust collector
, the two - point correlation function behaviour is strictly connected with cluster character of the fractal .
there is also no point to investigate the singularities at kissing limit because close packing is restricted by incoherent structure of the fractal .
the preliminary rsa simulation has been also performed for @xmath2 using general cantor set ( gcs ) as a collector .
the gcs dimension depends on the size of middle part removed in each iteration : @xmath65}\ ] ] where @xmath66 is the relative length of the removed fragment . for standard cantor set @xmath67 thus @xmath68 .
the diameter of the collector used during simulations was @xmath69 .
the dependence of @xmath70 on dimension value is shown in fig.[fig : qd06 ] . for @xmath2 .
dots represent data and dashed line is the ( [ eq : fit2 ] ) fit with @xmath56 , @xmath57 and @xmath58 - the same as presented in fig.[fig : qd16].,width=302 ] the data for @xmath71 are consistent with the fit ( [ eq : fit2 ] ) within an error range . for lower @xmath3 , when holes in a collector are bigger , the geometrical relations start to play the most important role .
the situation is somehow similar to the cantor dust case ; however for this case , it can be proved that random maximal coverage ratio depends strictly on the particle diameter to collector size ratio .
therefore , the possibility of getting reliable data for @xmath72 from rsa simulation is doubtful .
there is also no point in analysing the two - point correlation function because all collectors having @xmath2 are incoherent and the correlation behaviour will mainly express the structure of a specific collector .
the rsa on fractal collectors having @xmath0 was studied . for such systems ,
the adsorption kinetics obey the feder s law ( [ eq : feder ] ) .
moreover , the singularity of the two - point correlation function at kissing limit is logarithmic .
the decay seems to be superexponential , but the autocorrelations for low dimensional fractals reflect also collector structure .
it was one of the reasons why obtaining reliable results from simulation on collectors having dimension significantly lower than @xmath48 is not possible .
the maximal random coverage ratio for the whole range of collector dimensions is governed by the eq.([eq : fit2 ] ) , which is a simple extension to the rule ( [ eq : fit1 ] ) introduced and analytically supported for high dimensions .
this work was supported by grant mnisw/0013/h03/2010/70 .
10 s. kinge , m. crego - calama , d.n .
reinhoudt , chemphyschem * 9 * 1 20 ( 2008 ) .
p. pfeifer and d. avnir , j. chem .
phys . * 79 * , 3558 ( 1983 ) d. avnir , d. farin , and p. pfeifer , j. chem . phys . * 79 * , 3566 ( 1983 ) ; surf . sci . * 126 * , 569 ( 1983 ) ; nature * 30 * , 261 ( 1984 ) . e. basillais , comptes rendus de lacademie des sciences - serie iii , * 321 * ( 4 ) , 295 ( 1998 ) . m. m. khasanov and i. i. abyzbaev , j. eng .
. thermophys . * 61 * 6 , 1516 ( 1991 ) . j. h. conway and n. j. a. sloane ,
sphere packings , _ lattices and groups _ , springer - verlag , new york , 1998 .
c. a. rogers , _ packing and covering _ , cambridge university press , cambridge , 1964 .
cadilhe , v. privman , modern phys .
b * 18 * , 207 ( 2004 ) .
y.fan , j.k.percus , phys .
* 67 * 13 1677 ( 1991 )
. m. s. nazzarro , a. j. ramirez pastor , j. l. riccardo and v. pereyra , j. phys .
a : math . gen . * 30 * 1925 ( 1997 ) .
e. s. loscar , r. a. borzi , and e. v. albano , phys .
e * 68 * 041106 ( 2003 ) .
m. w. cole and n. s. holter , phys .
b * 33 * 8806 ( 1986 ) .
p. pfeifer , y. j. wu , m. w. cole , j. krim , phys .
* 62 * 1997 ( 1989 )
. n. v. brilliantov , yu . a. andrienko , p. l. krapivsky , and j. kurths , phys .
e * 58 * 3530 ( 1998 ) .
j. feder , j. theor .
87 , 237 ( 1980 ) .
r. h. swendsen , phys .
a * 24 * , 504 ( 1981 ) .
v. privman , j .- s .
wang , and p. nielaba , phys .
b * 43 * 3366 ( 1991 ) .
s. torquato , o.u .
uche , f.h .
stillinger , phys.rev.e * 74 * 061308 ( 2006 ) .
a. renyi , sel .
. prob . * 4 * , 203 ( 1963 ) .
inst . hung .
* 3 * , 109 , ( 1958 ) .
s. torquato and f. h. stillinger , j. phys .
b * 106 * , 8354 ( 2002 ) .
b. bonnier , d. boyer , and p. viot , j. phys .
a * 27 * , 3671 ( 1994 ) .
m. ciesla , j. barbasz , j. stat .
mech . 03015 ( 2012 ) .
z. adamczyk , j. barbasz , m. ciesla , langmuir , * 26 * 11934 ( 2010 ) ; langmuir , * 27 * 6868 ( 2011 ) . k. ball , int .
notices * 68 * , 217 ( 1992 ) . | irreversible adsorption of spheres on flat collectors having dimension @xmath0 is studied .
molecules are adsorbed on sierpinski s triangle and carpet like fractals ( @xmath1 ) , and on general cantor set ( @xmath2 ) .
adsorption process is modeled numerically using random sequential adsorption ( rsa ) algorithm .
the paper concentrates on measurement of fundamental properties of coverages , i.e. maximal random coverage ratio and density autocorrelation function , as well as rsa kinetics .
obtained results allow to improve phenomenological relation between maximal random coverage ratio and collector dimension .
moreover , simulations show that , in general , most of known dimensional properties of adsorbed monolayers are valid for non - integer dimensions . |
among interesting problems of modern cosmology the origin of large - scale magnetic fields is intensively attacked nowadays .
various mechanisms of the field generation at different stages of the universe evolution were proposed @xcite .
basically they are grounded on the idea of fermi , chandrasekhar and zeldovich that to have the present day galaxy magnetic fields of order @xmath5 correlated on a scale @xmath6mpc seed magnetic fields must be present in the early universe .
these fields had been frozen in a cosmic plasma and then amplified by some of the mechanisms of the field amplification .
one of the ways to produce seed fields is a spontaneous vacuum magnetization at high temperature @xmath7 @xcite .
actually , this is an extension of the savvidy model for the qcd vacuum @xcite , proposed already at @xmath8 and describing the creation of the abelian chromomagnetic fields due to a vacuum polarization , in case of nonzero temperature . at zero temperature
this field configuration is unstable because of the tachyonic mode in the gluon spectrum . at @xmath9 , the possibility of having strong temperature - dependent and stable magnetic fields
was discovered @xcite .
the field stabilization is ensured by the temperature and field dependent gluon magnetic mass .
another related field of interest is the deconfinement phase of qcd .
as it was realized recently , this is not the gas of free quarks and gluons , but a complicate interacting system of them .
this was discovered at rhic experiments @xcite and observed in either perturbative @xcite or nonperturbative @xcite investigations of the vacuum state with magnetic fields at high temperature . in refs .
@xcite the spontaneous creation of the chromomagnetic fields of order @xmath10 was observed in @xmath0- and @xmath11-gluodynamics within the one - loop plus daisy resummation accounted for . in ref .
@xcite the chromomagnetic condensate of same order was obtained in stochastic qcd vacuum model and method of dimensional reduction by comparison with lattice data . in refs .
@xcite the response of the vacuum to the influence of strong external fields at different temperatures has been investigated and it was shown that the confinement is restored by increasing the strength of the applied field .
these results stimulated the present investigation .
we are going to determine the spontaneous creation of magnetic fields in lattice simulations of @xmath0-gluodynamics .
in contrast to the problems in the external field , in the case of interest the field strength is a dynamical variable which values at different temperatures have to be determined by means of the minimization of the free energy .
this procedure is not a simple one as in continuum because the field strength on a lattice is quantized . to deal with this peculiarity
, we consider magnetic fluxes on a lattice as the main objects to be investigated .
the fluxes take continuous values , and therefore the minimization of the free energy in presence of magnetic field can be fulfilled in a usual way .
these speculations serve as an explanation of the strategy of our calculations .
one of the methods to introduce a magnetic flux on a lattice is to use the twisted boundary conditions ( t.b.c . )
@xcite . in this approach
the flux is a continuous quantity .
so , in what follows we consider the free energy @xmath12 with the magnetic flux @xmath13 on a lattice in the @xmath0-gluodynamics and calculate its values at different temperatures by means of monte carlo ( mc ) simulations .
we will show that the global minimum of @xmath12 is located at some non - zero value @xmath14 dependent on the temperature .
it means the spontaneous creation of the temperature - dependent magnetic fields in the deconfinement phase .
the paper is organized as follows . in sect .
2 some necessary information about the magnetic fluxes on a lattice is adduced . in sect .
3 the calculation details and the results are given .
section 4 is devoted to discussion .
in perturbation theory , the value of the macroscopic ( classical ) magnetic field generated inside a system is determined by the minimization of the free energy functional .
the interaction with the classical field is introduced by splitting the gauge field potential in two parts : @xmath15 , where @xmath16 describes a radiation field and @xmath17 corresponds to the constant magnetic field directed along the third axis . however , on a lattice , the direct detection of the spontaneously generated field strength by straightforward analysis of the configurations , which are produced in the mc simulations , seems to be problematic .
therefore , it is reasonable to follow the approach used in the continuum field theory .
first , let us write down the free energy density , @xmath18 \exp\{-s(u(\varphi))\}.\end{aligned}\ ] ] here , @xmath19 and @xmath20 are the partition function at finite and zero magnetic fluxes , respectively ; the link variable @xmath21 is the lattice analogue of the potential @xmath22 .
the free energy density relates to the effective action as follows , @xmath23 where @xmath24 and @xmath25 are the effective lattice actions with and without magnetic field , correspondingly . to detect the spontaneous creation of the field it is necessary to show that the free energy density has the global minimum at a non - zero magnetic flux , @xmath26 . in
what follows , we use the hypercubic lattice @xmath27 ( @xmath28 ) with the hypertorus geometry ; @xmath29 and @xmath30 are the temporal and the spatial sizes of the lattice , respectively . in the limit of @xmath31 the temporal size @xmath29 is related to physical temperature .
the one - plaquette action of the @xmath0 lattice gauge theory can be written as @xmath32;\\\label{plaq } u_{\mu\nu}(x)=u_\mu(x)u_\nu(x+a\hat{\mu})u_\mu^\dagger(x+a\hat{\nu})u_\nu^\dagger(x),\end{aligned}\ ] ] where @xmath33 is the lattice coupling constant , @xmath34 is the bare coupling , @xmath35 is the link variable located on the link leaving the lattice site @xmath36 in the @xmath37 direction , @xmath38 is the ordered product of the link variables .
the effective action @xmath39 in ( [ enactions ] ) is the wilson action averaged over the boltzmann configurations , produced in the mc simulations .
the lattice variable @xmath35 can be decomposed in terms of the unity , @xmath40 , and pauli , @xmath41 matrices in the color space , @xmath42 the four components @xmath43 are subjected to the normalization condition @xmath44 .
hence , only three components are independent .
since the spontaneously generated magnetic field is to be the abelian one , the abelian parametrization of the lattice variables is used to introduce the magnetic field , @xmath45 where the angular variables are changed in the following ranges @xmath46 , @xmath47 . the abelian part of the lattice variables is represented by the diagonal components of the matrix and the condensate abelian magnetic field influences the field @xmath48 , only .
the second important task is to incorporate the magnetic flux in this formalism .
the most natural way was proposed by t hooft @xcite . in his approach , the constant homogeneous external flux @xmath13 in the third spatial direction can be introduced by applying the following t.b.c .
: @xmath49 it could be seen , the edge links in all directions are identified as usual periodic boundary conditions except for the links in the second spatial direction , for which the additional phase @xmath13 is added ( fig .
1 ) . in the continuum limit , such t.b.c . settle the magnetic field with the potential @xmath50 .
the magnetic flux @xmath13 is measured in angular units and can take a value from @xmath51 to @xmath52
. the lattice variables ( in the abelian parametrization ) in the presence of the magnetic flux @xmath13 are @xmath53 where @xmath54 for the edge links at @xmath55 with @xmath56 and @xmath57 for other links .
the total flux through the plane spanned by the plaquettes @xmath58 , which affects the edge links at @xmath55 with @xmath56 , is @xmath59 eq .
( [ totflux ] ) is the lattice analogue of the flux in the continuum : @xmath60 in this approach the variable @xmath13 describes a flux through the whole lattice plane , not just through an elementary plaquette .
the t.b.c . for the components ( [ paramphi ] ) ,
@xmath61 read @xmath62 the relations ( [ tbc2 ] ) and ( [ tbc3 ] ) have been implemented into the kernel of the mc procedure in order to produce the configurations with the magnetic flux @xmath13 . in this case the flux @xmath13 is accounted for in obtaining a boltzmann ensemble at each mc iteration .
the mc simulations are carried out by means of the heat bath method .
the lattices @xmath63 , @xmath2 and @xmath64 at @xmath65 , @xmath66 are considered .
these values of the coupling constant correspond to the deconfinement phase and perturbative regime . to thermalize the system ,
200 - 500 iterations are fulfilled . at each working iteration
, the plaquette value ( [ plaq ] ) is averaged over the whole lattice leading to the wilson action ( [ wilson ] ) .
then the effective action is calculated by averaging over the 1000 - 5000 working iterations . by setting a set of magnetic fluxes @xmath13 in the mc simulations we obtain the corresponding set of values of the effective action .
the value of the condensed magnetic flux @xmath14 is obtained as the result of the minimization of the free energy density ( [ enactions ] ) .
the spontaneous generation of magnetic field is the effect of order @xmath67 @xcite .
the results of mc simulations show the comparably large dispersion .
so , the large amount of the mc data is collected and the standard @xmath4-method for the analysis of data is applied to determine the effect .
we consider the results of the mc simulations as observed `` experimental data '' .
the effective action depends smoothly on the flux @xmath13 in the region @xmath68 .
therefore , the free energy density can be fitted by the quadratic function of the flux @xmath13 , @xmath69 this choice is motivated also by the results obtained already in continuum field theory @xcite where it was determined that free energy has a global minimum at @xmath70 .
the parametrization ( [ parfit ] ) is the most reasonable in this case .
it is based on the effective action accounting for the one - loop plus daisy diagrams @xcite , @xmath71^{3/2},\end{aligned}\ ] ] having @xmath72 and @xmath73 orders in coupling constant . here , @xmath74 is field strength ( flux @xmath75 ) , @xmath7 is the temperature , @xmath37 is the normalization point , @xmath76 is the zero zero component of the gluon polarazation operator calculated in the external field at the finite temperature and taken at zero momentum .
the value of @xmath77 , which was used , corresponds to a deep perturbation regime .
so , a comparison with perturbation results is reasonable .
the systematic errors in fitting function ( [ parfit ] ) could come from not taking into account the high - order diagrams in ( [ diagr ] ) .
however , as it is well known @xcite , the lack of an expansion parameter at finite temperature starts from the three - loop diagram contributions that is of @xmath78 order and could not remove an effect derived in @xmath72 and @xmath79 orders .
as the finite - size effects are concerned , in the present investigation we just made calculations for two lattices @xmath63 and @xmath2 and have derived the same results for the @xmath14 ( as it will be seen below ) .
a more detailed investigation of this issue requires much more computer resources , which were limited .
there are 3 unknown parameters , @xmath80 , @xmath81 and @xmath14 in eq.([ffit ] ) .
the parameter @xmath14 denotes the minimum position of the free energy , whereas the @xmath80 and @xmath81 are the free energy density at the minimum and the curvature of the free energy function , correspondingly .
the value @xmath14 is obtained as the result of the minimization of the @xmath4-function @xmath82 where @xmath83 is the array of the set fluxes and @xmath84 is the data dispersion .
it can be obtained by collecting the data into the bins ( as a function of flux ) , @xmath85 where @xmath86 is the number of points in the considered bin , @xmath87 is the mean value of free energy density in the considered bin .
as it is determined in the data analysis , the dispersion is independent of the magnetic flux values @xmath13 .
the deviation of @xmath14 from zero indicates the presence of spontaneously generated field .
-fit of the free energy density on lattice @xmath63 ( left figure ) and @xmath2 ( right figure ) for @xmath65 ( grey regions describe the @xmath88 and @xmath89 , at the 95% c.l . , correspondingly).,scaledwidth=40.0% ] ' '' '' & @xmath63 & @xmath2 & @xmath90 + ' '' '' @xmath65 & @xmath91 & @xmath92 & @xmath93 + ' '' '' @xmath94 & @xmath95 & & + the fit results are given in the table 1 .
as one can see , @xmath14 demonstrates the @xmath96-deviation from zero .
the dependence of @xmath14 on the temperature is also in accordance with the results known in perturbation theory : the increase in temperature results in the increase of the field strength @xcite .
the fits for the lattices @xmath63 and @xmath2 at @xmath65 are shown in fig .
the maximum - likelihood estimate of @xmath12 by the whole data set is shown as the solid curve .
in addition , all @xmath13 values are divided into 15 bins .
the mean values and the 95% confidence intervals are presented as points for each bin .
the first 9 bins contain about 600 - 2000 points per bin .
the large number of points in the bins allow to find the free energy @xmath97 with the accuracy which substantively exceeds the dispersion , @xmath98 .
it makes possible to detect the effect of interest .
as it is also seen , the maximum - likelihood estimate of @xmath12 is in a good accordance with the bins pointed , because the solid line is located in the 95% confidence intervals of all bins . the 95% c.l .
area of the parameters @xmath80 ( @xmath81 for the right figure ) and @xmath14 is represented in fig .
the black cross marks the position of the maximum - likelihood values of @xmath80 ( @xmath81 for the right figure ) and @xmath14 .
it can be seen that the flux is positive determined .
the 95% c.l .
area becomes more symmetric with the center at the @xmath80 , @xmath81 and @xmath14 when the statistics is increasing .
this also confirms the results of the fitting . and
@xmath14 , determining the free energy density dependence on the flux @xmath14 on lattice @xmath63 for @xmath65 ( left figure ) .
the 95% c.l .
area for the parameters @xmath80 and @xmath81 , determining the free energy density dependence on the flux @xmath14 on lattice @xmath2 for @xmath65 ( right figure).,title="fig:",scaledwidth=30.0% ] and @xmath14 , determining the free energy density dependence on the flux @xmath14 on lattice @xmath63 for @xmath65 ( left figure ) .
the 95% c.l .
area for the parameters @xmath80 and @xmath81 , determining the free energy density dependence on the flux @xmath14 on lattice @xmath2 for @xmath65 ( right figure).,title="fig:",scaledwidth=30.0% ]
the main conclusion from the results obtained is that the spontaneously created temperature - dependent chromomagnetic field is present in the deconfinement phase of qcd .
this supports the results derived already in the continuum quantum field theory @xcite and in lattice data analysis @xcite .
let us first discuss the stability of the magnetic field at high temperature .
it was observed in refs .
@xcite that the stabilization happens due to the gluon magnetic mass calculated from the one - loop polarization operator in the field at temperature .
this mass has the order @xmath99 as it should be because the chromomagnetic field is of order @xmath100 @xcite .
the stabilization is a nontrivial fact that , in principle , could be changed when the higher order feynman diagrams to be accounted for .
now we see that the stabilization of the field really takes place .
our approach based on the joining of calculation of the free energy functional and the consequent statistical analysis of its minimum positions at various temperatures and flux values .
this overcomes the difficulties peculiar to the description of the field on a lattice .
here we mean that the field strength on a lattice is quantized and therefore a nontrivial tuning of the coupling constant , temperature and field strength values has to be done in order to determine the spontaneously created magnetic field .
we also would like to note that in the present paper the flux dependence on temperature remains not investigated in details .
this is because of the small lattice size considered .
that restricts the number of points permissible to study .
however , at this stage we have determined the effect of interest as a whole . even at the small lattice
, one needs to take into consideration thousands points of free energy ( that corresponds to an analysis of 5 - 10 millions mc configurations for different lattices ) to determine the flux value @xmath14 at the @xmath101 c.l . in case of larger lattices this number and corresponding computer resources should be increased considerably .
this problem is left for the future .
as we mentioned in section 2 , the finite - size effects have not been investigated in detail . however , these effects are important near the phase transition temperature .
they make difficult to distinguish a first - order phase transition from a second - order one . in our case ,
the temperature is far from @xmath102 .
the fact that external field penetrates the coulomb phase is well known @xcite , so the only really new thing is that this field is spontaneously created .
it was first observed in continuum @xcite , where the field strength of order @xmath103 in coupling constant was established .
finite - size effects are not able to remove this result .
the values @xmath14 obtained on the lattices @xmath63 and @xmath2 ( see the table 1 ) are in a good agrement with each other , within the statistical errors at @xmath101 c.l .
one could speculate that the lattice sizes @xmath63 and @xmath2 are not sufficient
. however , these lattice sizes were used in the refs . @xcite .
the main aim of present paper is to show a possibility of spontaneous generation of chromomagnetic field at high temperature in lattice simulation , which was investigated already by perturbative methods @xcite .
it is interesting to compare our results with that of in ref .
@xcite where the response of the vacuum on the external field was investigated .
these authors have observed in lattice simulations for the @xmath0- and @xmath11-gluodynamics that the external field is completely screened by the vacuum at low temperatures , as it should be in the confinement phase . with the temperature increase ,
the field penetrates into the vacuum and , moreover , increase in temperature results in existing more strong external fields in the vacuum . on the other hand ,
increase in the applied external field strength leads to the decreasing of the deconfinement temperature .
these interesting properties are closely related to the studies in the present work .
actually , we have also investigated the vacuum properties as an external field problem when the field is described in terms of fluxes .
this was the first step of the calculations .
the next step was the statistical analysis of the minimum position of free energy , in order to determine the spontaneous creation of the field .
in fact , at the first step we reproduced the results of refs .
@xcite ( in terms of fluxes ) .
note that the present investigations also correspond to the case of the early universe .
they support our previous results on the magnetic field generation in the standard model @xcite and in the minimal supersymmetric standard model @xcite .
as it was discussed by pollock @xcite , the field generated by this mechanism at the planck era might serve as a seed field to produce the present day magnetic fields in galaxies .
we would like to conclude with the note that the deconfinement phase of gauge theories is a very interesting object to study .
the temperature dependent magnetic fields , which are present in this state , influence various processes that should be taken into consideration to have an adequate concept about them .
the authors would like to express sincere gratitude to michael ilgenfritz for his kind attention and help at each stage of the work .
we also thank alexey gulov for numerous useful discussions .
one of us ( vd ) is indebted for hospitality to ictp ( trieste ) , where the final part of the work has been done .
20 d. grasso , h. rubinstein , phys .
( 2001 ) 163 .
k. enqvist , p. olesen , phys .
b 329 ( 1994 ) 195 .
a. starinets , a. vshivtsev , v. zhukovsky , phys .
b 322 ( 1994 ) 403 .
v. skalozub , m. bordag , nucl .
b 576 ( 2000 ) 430 .
m. pollock , int . j. mod
d 12 ( 2003 ) 1289 .
g. savvidy , phys .
b 71 ( 1977 ) 133 .
star collaboration , j. adams , et al .
phys . a 757 ( 2005 ) 102 .
v. skalozub , a. strelchenko , eur .
j. c 40 ( 2005 ) 121 .
n. agasian , phys .
b 562 ( 2003 ) 257 .
p. cea , l. cosmai , phys .
d 60 ( 1999 ) 094506 ; + p. cea , l. cosmai , jhep 0508 ( 2005 ) 079 ; + p. cea , l. cosmai , hep - lat/0101017 .
g. t hooft , nucl .
b 153 ( 1979 ) 141 .
v. skalozub , a. strelchenko , eur .
j. c 33 ( 2004 ) 105
. v. demchik , v. skalozub , eur .
j. c 25 ( 2002 ) 291 .
v. demchik , v. skalozub , eur .
j. c 27 ( 2003 ) 601 .
a. linde , rept .
42 ( 1979 ) 389
. m. vettorazzo , ph . de forcrand , nucl . phys .
b 686 ( 2004 ) 85 . t.a .
degrand , d. toussaint , phys .
d 25 ( 1982 ) 526 ; + j. engels , f. karsch , h. satz , phys .
b 101 ( 1981 ) 89 ; + j. ambjorn , v. mitrjushkin , v. bornyakov , a. zadorozhnyi , phys .
b 225 ( 1989 ) 153 ; + j. fingberg , urs m. heller , v.k .
mitrjushkin , nucl .
b 435 ( 1995 ) 311 . | the spontaneous generation of the chromomagnetic field at high temperature is investigated in a lattice formulation of the @xmath0-gluodynamics .
the procedure of studying this phenomenon is developed .
the monte carlo simulations of the free energy on the lattices @xmath1 , @xmath2 and @xmath3 at various temperatures are carried out .
the creation of the field is indicated by means of the @xmath4-analysis of the data set accumulating 5 - 10 millions mc configurations .
a comparison with the results of other approaches is done . |
here we present the slave - fermion mean - field results for the @xmath59 model with @xmath33 , that allow us to interpret the physical origin of the kinetic antiferromagnetism found numerically with the dmrg method .
first , we briefly introduce the slave - fermion mean - field treatment @xcite of the @xmath59 model , @xmath104 where the sum @xmath105 is over all the @xmath65 sites of the cluster and all the @xmath80 neighbors connected with @xmath65 by the exchange interaction @xmath106 and hopping term @xmath107 the @xmath59 model is the strong coupling @xmath108 limit of the hubbard model , with @xmath109 , and it is written in terms of the projected electronic operators , @xmath110 with @xmath111 forbidding the electronic double occupancy at each site . for @xmath3 , @xmath112 , and both models are identical .
the projected electronic operators can be represented in term of schwinger bosons @xmath113 and spinless fermions @xmath114 @xmath115 this replacement , together with the constraint @xmath116 at each site @xmath65 , is a faithful representation of the original fermi algebra .
the spin can be written in terms of the schwinger bosons as @xmath117 , where @xmath118 is the vector of pauli matrices , while the slave fermion @xmath119 corresponds to the charge ( hole ) degree of freedom .
we replace ( [ cproj ] ) in the @xmath59 hamiltonian ( [ tj ] ) and we obtain @xmath120 , \label{tjslave}\ ] ] where we have defined the singlet @xmath6 operators @xmath121 and @xmath122 @xcite , while @xmath123 .
@xmath124 ( @xmath125 ) is related with the antiferromagnetic ( ferromagnetic ) correlations between sites @xmath65 and @xmath126 , while @xmath127 represents the probability that a hole hopes from @xmath65 to @xmath126 .
hamiltonian ( [ tjslave ] ) involves quartic terms in slave operators , and thus we appeal to a mean - field decoupling @xcite in order to solve it approximately . also , we approximate the local constraint ( [ const ] ) by its average over all the lattice , enforced by a lagrange multiplier @xmath128 .
the resulting mean - field hamiltonian is @xmath129 to obtain this expression we replace the bilinear operators @xmath130 , @xmath131 and @xmath132 by its ( real ) averages , assuming a translational invariant solution . also , we introduce the chemical potential @xmath133 to control the hole doping @xmath134 . after transforming to the fourier space , we get @xmath135 + { \rm cte . } , \label{tjk}\ ] ] where the hole dispersion is @xmath136 while @xmath137 and @xmath138 now we have a quadratic hamiltonian ( [ tjk ] ) , that can be straightforwardly diagonalized by means of a bogoliubov transformation of the schwinger bosons to new bosonic operators @xmath139 s @xcite : @xmath140 where the boson dispersion is @xmath141 finally , the mean field parameters @xmath142 , @xmath143 and @xmath144 can be computed in a self consistent way @xcite .
the mean - field hamiltonian ( [ mf ] ) approximates the complicated @xmath59 model by two one - body terms : a fermionic tight - binding - like one , describing the motion of holes , and a bosonic term , corresponding to the spin fluctuations of the system .
it should be noticed that both terms are interdependent , as the hole dispersion depends on the ferromagnetic parameter @xmath82 while the bosonic dispersion depends on @xmath144 , providing a coupling between the charge and spin degrees of freedom .
once we obtain the mean field hamiltonian , it is a simple task to calculate the ground state physical properties of the system as a function of @xmath73 , @xmath145 and hole doping , @xmath146 in particular , the existence of magnetic order is manifested in the bose - einstein condensation of the bosons at @xmath147 , where @xmath148 is the magnetic wave vector @xcite . in fig . 2 of the manuscript we show the mean field local magnetization of the triangular @xmath59 model as a function of @xmath149 in order to illustrate the ordering effect of kinetic antiferromagnetism in this lattice .
in a future publication @xcite , we will present a detailed study of the mean field phase diagram of the @xmath59 model on the triangular and square lattices . hereafter
, we report the mean field results for the special case of interest in this work , that is , @xmath33 .
nel , ferromagnetic , and the collinear stripe order , for @xmath33 and @xmath4 in the triangular lattice.,width=264 ] for @xmath33 , the @xmath150 factors vanish identically , consequently hamiltonian ( [ tjk ] ) is already diagonal and there is no need of a bogoliubov transformation . instead
, it is imperative to condense the schwinger bosons , in order to satisfy the averaged constraint ( [ const ] ) . considering bose condensation at different momenta @xmath151
, we can calculate the energy of magnetic phases ( if the phases are locally stable , that is , they satisfy the self - consistent mean - field equations ) characterized by different magnetic wave vectors @xmath148 s . in particular , for the triangular lattice and @xmath85 , we evaluate the energy of the ferromagnetic phase ( @xmath152 ) , the @xmath2 nel order ( @xmath153 ) , and a collinear stripe order ( @xmath154 ) , that is favorable for intermediate doping and @xmath95 @xcite . in fig .
[ enertpos ] we show the energy _ per site _ of the three phases as a function of hole doping , calculated for a large cluster ( @xmath155 sites ) that represents the thermodynamic limit .
it can be seen that , at the mean - field level , the ferromagnetic phase is the ground state for all the doping range considered , although the 120@xmath89 nel order has a very close energy for doping @xmath156 .
the collinear stripe phase always has a much larger energy , and so it is not stabilized by the kinetic energy alone .
it should be stressed that after an exhaustive search for solutions of the mean - field equations for finite doping , we have always found that the ferromagnetic phase is the ground state . in the case of only one hole doped away half - filling , we will show in the next section that the ferromagnetic phase is degenerate with the @xmath2 nel order .
the energy of both phases is @xmath157 ; while this is the exact ferromagnetic phase energy , it is clearly rather above the exact 120@xmath89 nel phase energy found by our dmrg calculations and exact diagonalization @xcite .
it is worthy to note that , once @xmath158 , the 120@xmath89 nel state is strongly favored over the ferromagnetic one @xcite , because the exchange interaction select the 120@xmath89 magnetic pattern in the triangular lattice @xcite . for @xmath159
, it can be seen in fig .
[ enertneg ] that the ferromagnetic phase is also the ground state for all doping , but now its energy is well separated from the energies of the other phases considered ( 120@xmath89 nel phase and a representative spiral phase , with magnetic wave vector @xmath160 ) . for negative @xmath11 nagaoka s theorem is valid , and it is believed that ferromagnetism is very stable for further doping @xcite .
nel , ferromagnetic , and spiral @xmath160 phases , for @xmath33 and @xmath161 in the triangular lattice.,width=264 ] so , for @xmath85 and @xmath162 the mean - field approximation does not capture that @xmath2 nel order is the unique ground state , as it is degenerate with the ferromagnetic phase .
this failure has to do with the mean - field character of the approach , in which we approximate the actual many - body problem by one - body terms . from previous experience @xcite
, we expect that the inclusion of gaussian fluctuations above the mean - field solutions will put the slave - fermion predictions in consonance with our dmrg and exact diagonalization @xcite results , lowering the 120@xmath89 nel energy while leaving unchanged the already exact ferromagnetic energy .
for the @xmath59 model on the square lattice with first and second neighbor hoppings , with @xmath112 and @xmath163 , we obtained analogous results to the triangular lattice . for @xmath40 ,
the ferromagnetic phase is clearly the ground state , in accordance with the nagaoka s theorem . for @xmath164
although the ferromagnetic phase is the ground state for all doping considered ( @xmath165 ) , its energy is very close to the energy of the usual @xmath166 nel state , being both states degenerate for one hole doped away half - filling .
again , this result does not agree with the dmrg prediction of a unique antiferromagnetic ground state .
although the mean - field theory fails to obtain the correct antiferromagnetic ground states for @xmath33 , in the next section we argue that it is a very helpful approach to elucidate the physical mechanism behind the phenomenon of kinetic antiferromagnetism .
in this section we limit ourselves to the case of one hole doped , relevant for our work .
we analyze why the 120@xmath89 nel order ( @xmath166 ) is one of the ground states , along with the ferromagnetic one , of the mean field hamiltonian ( [ tjk ] ) for the triangular lattice ( square lattice ) .
the answer lies in the release of the hole kinetic energy frustration thanks to the hole motion on the antiferromagnetic background , and we believe that this mechanism can be used to interpret the kinetic antiferromagnetism . in the one hole case , the ground state energy of the system corresponds to the energy of the bottom of the hole dispersion , @xmath74 ( [ edisp ] ) .
this energy dispersion is a tight - binding like , with the hopping terms @xmath73 renormalized by the ferromagnetic mean - field parameter @xmath75 @xmath76 . on one hand , this renormalization can induce a spin - berry phase , while , on the other hand , it could lead to the vanishing of the hopping amplitude between certain neighbor sites . although , the mechanism is the same for both lattices , the details differ , so we analyze each case separately . for @xmath162
the analysis of the mean - field equations shows that the @xmath82 parameters take their classical values , @xmath172 @xcite .
so , for the ferromagnetic phase , @xmath152 , the hole dispersion takes the form @xmath173,\ ] ] while for the 120@xmath89 nel order , @xmath174 , @xmath175.\ ] ] the minimum and maximum ( maximum and minimum ) of the ferromagnetic energy dispersion occur at the momenta @xmath176 and @xmath177 , respectively , for @xmath4 ( @xmath159 ) : @xmath178 as a consequence , the hole motion is frustrated ( unfrustrated ) for @xmath85 ( @xmath159 ) . in the 120@xmath89 nel phase , on one hand , the hole dispersion bandwidth is halved .
this results from the fact that the effective hopping @xmath81 between sites @xmath65 and @xmath126 is proportional , through the modulus of @xmath82 , to the overlap of the spin wavefunctions pointing in the direction of the local magnetization at each site , similar to what happens in the double exchange mechanism @xcite , @xmath183 where @xmath79 is the angle between spins at sites @xmath65 and @xmath126 . on the other hand
, the negative sign acquired by the @xmath184 parameter turns upside down the hole dispersion with respect to the ferromagnetic case . as the energy of @xmath59 model , with @xmath33 and one hole
doped away half - filling , coincides with the bottom of the hole dispersion energy , for @xmath159 the ferromagnetic phase is the ground state , with a significantly lower energy than the 120@xmath89 nel state .
the mean - field approximation gets the exact ground state energy , @xmath185 , and it satisfies nagaoka s theorem . as it is shown in fig .
( [ enertneg ] ) ferromagnetism clearly prevails for further doping .
in contrast , for @xmath93 the hole motion in the ferromagnetic state is frustrated , while the kinetic energy frustration is released in the 120@xmath89 state thanks to the negative @xmath184 parameter . at the mean - field level
this release gives rise to the degeneracy of the ferromagnetic and the 120@xmath89 nel phases . while the mean - field ferromagnetic energy is the exact one ( @xmath186 ) , the mean - field energy of the 120@xmath89 state is approximate , and it is necessary to use more sophisticated techniques beyond mean - field , like the dmrg computations we perform in our work or the exact diagonalization by haerter and shastry @xcite , to capture the stabilization of the 120@xmath89 phase over the ferromagnetic one .
it is worth to notice that the energy difference between the ferromagnetic and the 120@xmath89 phases computed by dmrg , @xmath187 , is relatively large ( @xmath188 of the ferromagnetic energy ) , indicating that many - body effects beyond mean - field are really important . in spite of this discrepancy of the mean - field approximation
, we believe that it is a very helpful approach because we can extract from it the main ingredient of the physical mechanism that gives rise to the kinetic antiferromagnetism in these extremely correlated systems : in kinetically frustrated situations when nagaoka s theorem is not valid , the kinetic frustration can be released if the hole moves in certain antiferromagnetic backgrounds , lowering its energy with respect to the ferromagnetic state ( when quantum fluctuations beyond mean - field are considered ) .
the sum of the ( complex ) phases of the @xmath82 parameters around a closed loop is the spin - berry phase acquired by the hole when it runs through the loop , and it is also half the solid angle subtended by the spins on the loop @xcite . while in the ferromagnetic phase ( like in any collinear phase ) the subtended solid angle vanishes , in the 120@xmath89 nel phase this angle is @xmath189 for each unit triangle , and its associated spin - berry phase is @xmath190 ( related to the negative sign of @xmath184 @xcite ) .
so , we can summarize that in the triangular lattice the mechanism of kinetic antiferromagnetism consists in the release of the kinetic frustration by means of the spin - berry phase acquired by the hole as it moves along the 120@xmath89 magnetic pattern . in the square lattice with hopping terms to first and second neighbors
, the hole dispersion ( [ edisp ] ) is @xmath191 + 4t_2\left[b_3 \cos\left(k_x+k_y\right)+ b_4 \cos\left(-k_x+k_y\right)\right],\ ] ] where the @xmath168 s parameters correspond to the first neighbors @xmath169 and @xmath192 , and to the second neighbors , @xmath193 and @xmath194 .
as the @xmath12 sign is irrelevant , we take @xmath197 in the following . for @xmath198 ,
the minimum ( maximum ) of the ferromagnetic energy dispersion occurs at the momenta , @xmath199 or @xmath39 ( @xmath200 ) for @xmath201 @xmath202 while for @xmath40 the minimum ( maximum ) occurs at the momenta @xmath203 ( @xmath199 or @xmath39 ) : @xmath204 for the @xmath166 nel phase , the first neighbor @xmath82 parameters vanish identically because of the antiparallel orientation of the spins , while @xmath206 and @xmath207 .
therefore , the hole dispersion is @xmath208 as it is well known for this kind of mean - field approximations , the holes can only propagate along the ferromagnetic sublattices .
the minimum and maximum ( maximum and minimum ) of the hole dispersion now are located at momenta @xmath209 and @xmath210 , respectively , for @xmath88 ( @xmath40 ) , with @xmath211 for @xmath40 , the unique ground state of the @xmath59 model , with @xmath33 and one - hole doped away half - filling , is the ferromagnetic phase , with the optimal kinetic energy @xmath212 well separated from the nel phase energy ( @xmath213 ) .
as in the triangular case , the mean - field approximation gives the exact ground state energy and satisfies nagaoka s theorem .
for @xmath88 the ferromagnetic and nel states are degenerate at the mean field level ( @xmath214 ) . as in the triangular case ,
the ferromagnetic phase mean - field energy is the exact one , while the mean - field nel energy is approximate .
our dmrg computations shows that the nel phase is the true ground state , with an energy @xmath215 well below the mean - field value . in the square lattice model
, we can also conclude that the origin of the kinetic antiferromagnetism is intimately related with the release of the kinetic energy frustration in the nel phase . in this case ,
the kinetic frustration release is caused by the vanishing of the frustrating first neighbor hopping amplitudes in the @xmath166 nel state , and not by the appearance of a non - trivial spin - berry phase like in the triangular lattice .
so far , we have found two examples
where if the @xmath59 model with @xmath33 equivalently , the @xmath3 hubbard model is kinetically frustrated , there is a ground state degeneracy between a determined antiferromagnetic state and the ferromagnetic one at the mean field level .
the selected antiferromagnetic state is the one which allows the release of the kinetic frustration .
our dmrg calculations allow us to affirm that in these cases , the antiferromagnetic state is the true ground state of the model , indicating that the phenomenon of kinetic antiferromagnetism seems to be quite ubiquitous in kinetically frustrated systems . here , we briefly extend this kind of reasoning to other systems .
* we consider the mean - field analysis for the square lattice model with arbitrary @xmath216 .
we find that for any negative @xmath13 there is no kinetic frustration and , therefore , the nagaoka ferromagnetic state is the unique ground state . on the other hand , for positive @xmath13
there is kinetic frustration . if @xmath217 ( @xmath12 sign is irrelevant ) the usual @xmath5 nel and the ferromagnetic states are degenerate , while if @xmath218 the ferromagnetic phase is the only ground state , with an energy difference relative to the nel phase , @xmath219 based in the previous examples , we expect that , in an exact calculation , the @xmath5 nel phase will be the ( only ) ground state for @xmath217 .
we can push further our prediction : the ferromagnetic energy is the exact one , while the antiferromagnetic energy can still decrease if we go beyond the mean - field approximation , and taking into account that previous dmrg computations give energy differences between the antiferromagnetic and the ferromagnetic states of the order of @xmath220 , we propose that the nel state will be the ground state as long as @xmath221 , that is , for @xmath222 .
our preliminary dmrg results confirm this prediction , as we find that for @xmath223 the ground state has @xmath5 nel order , while for @xmath224 the ground state is ferromagnetic . *
the anisotropic triangular lattice has hopping @xmath11 along the @xmath225 vectors , and @xmath226 along the @xmath227 and @xmath228 vectors . for negative @xmath11 , independently of the @xmath226 sign , the system is kinetically unfrustrated and the mean - field approach yields a ferromagnetic ground state , with an energy well separated from the antiferromagnetic phases . for positive
@xmath11 , the kinetic energy is frustrated , and we find two different mean - field regimes : _ i _ ) for @xmath229 the ferromagnetic state is degenerate with a spiral phase characterized by a wave vector @xmath230 with @xmath231 going continuously from 0 ( @xmath232 ) to @xmath189 ( @xmath233 ) ; _ ii _ ) for @xmath234 the ferromagnetic phase is the only ground state . drawing on arguments similar to those put forward in the @xmath91 square lattice model
, we expect a kinetic antiferromagnetic ground state for @xmath235 and a ground state with ferromagnetic order for @xmath236 .
preliminary dmrg results confirm again this prediction , but the wave vector of the antiferromagnetic phases are different from the mean - field ones .
dmrg yield a 120@xmath89 nel order for @xmath237 while @xmath238 for @xmath239 with @xmath231 going from @xmath240 ( @xmath241 ) to @xmath242 ( @xmath243 ) .
the prevalence of the 120@xmath89 nel phase would indicate that this order optimize the release of the kinetic energy frustration , but more work is needed to assert this affirmation .
* finally we consider the @xmath3 hubbard model on the kagom lattice . for negative @xmath11
there is no kinetic frustration , and the mean - field recovers the exact nagaoka ground state , while for positive @xmath11 we find that the ferromagnetic phase is degenerate with mean - field 120@xmath89 magnetic orders , like the @xmath244 and @xmath245 states @xcite .
we speculate that , at the mean - field level , any state that obeys the condition of zero total spin at each triangle , that is , any classical ground state of the kagom heisenberg model , will be degenerate in energy with the ferromagnetic phase .
we leave for future investigations the very interesting analysis of kinetic antiferromagnetism on the kagom lattice .
99 c. jayaprakash , h.r .
krishnamurthy , and s. sarker , phys .
b * 40 * , 2610 ( 1989 ) .
ceccatto , c.j .
gazza , and a.e .
trumper , phys .
b * 47 * , 12329 ( 1993 ) .
a. mezio , c.n .
sposetti , l.o .
manuel , and a.e .
trumper , j. phys . : condens .
matter * 25 * , 465602 ( 2013 ) .
s. sarker , c. jayaprakash , h.r .
krishnamurthy , and m. ma , phys .
b * 40 * , 5028 ( 1989 ) . c. n. sposetti , a. mezio , a. e. trumper , c. j. gazza , and l. o. manuel ( in preparation ) . c. weber , a. laeuchli , f. mila , and t. giamarchi , phys .
b * 73 * , 014519 ( 2006 ) .
j. o. haerter and b. s. shastry , phys .
lett . * 95 * , 087202 ( 2005 ) .
manuel , a.e .
trumper , and h.a .
ceccatto , phys .
b * 57 * , 8348 ( 1998 ) . t. koretsune and m. ogata , phys . rev
89 * , 116401 ( 2002 ) .
trumper , l.o .
manuel , c.j .
gazza , and h.a .
ceccatto , phys .
lett . * 78 * , 2216 ( 1997 ) .
manuel and h.a .
ceccatto , phys .
b * 60 * , 9489 ( 1999 ) . in a finite cluster ,
the classical value is @xmath246 here we are considering the thermodynamic limit of this expression . c. zener , phys .
rev . * 82 * , 403 ( 1951 ) ; p.w . anderson and h. hasegawa , phys .
rev . * 100 * , 675 ( 1955 ) .
g. misguich and c. lhuillier , in _ frustrated spin systems _ , edited by h.t .
diep ( world scientific , 2005 ) . in our case the sign of @xmath184 determines the spin - berry phase .
we could have assigned this sign to any other @xmath82 parameter , chosing another gauge - related mean - field solution , see f. wang and a. vishwanath , phys .
b * 74 * , 174423 ( 2006 ) .
harris , c. kallin , and a.j .
berlinsky , phys .
b * 45 * , 2899 ( 1992 ) . | we study the infinite @xmath0 hubbard model with one hole doped away half - filling , in triangular and square lattices with frustrated hoppings that invalidate nagaoka s theorem , by means of the density matrix renormalization group .
we find that these kinetically frustrated models have antiferromagnetic ground states with classical local magnetization in the thermodynamic limit .
we identify the mechanism of this kinetic antiferromagnetism with the release of the kinetic energy frustration as the hole moves in the established antiferromagnetic background .
this release can occurs in two different ways : by a non - trivial spin - berry phase acquired by the hole or by the effective vanishing of the hopping amplitude along the frustrating loops .
itinerant magnetism has proved to be an elusive subject in condensed matter physics , since itinerant and localized aspects of electrons need to be taken into account on equal footing .
the single - band hubbard model , originally proposed to describe metallic ferromagnetism @xcite , has also been associated with antiferromagnetism of kinetic exchange origin close to half - filling . while _ virtual _ kinetic processes favor antiferromagnetism , it is a rule of thumb to link _ real _ kinetic processes with ferromagnetism @xcite .
however , there exist only few exact results ensuring the existence of itinerant ferromagnetism @xcite . among them ,
the most renowned is nagaoka s theorem @xcite , which assert that the saturated ferromagnetic state is the unique ground state when one hole is doped on the half - filled hubbard model with infinite @xmath0 coulomb repulsion .
furthermore , a connectivity condition must be fulfilled for the validity of nagaoka s theorem : the sign of the hopping amplitudes around the smallest closed loop of the lattice must be positive , otherwise the hole kinetic energy will be frustrated and the saturated ferromagnetic state will no longer be the ground state . kinetic energy frustration is a quantum mechanical phenomenon without classical analog , easily understood in certain tight - binding models where an electron can not gain the full kinetic energy @xmath1 due to quantum interferences @xcite .
this kind of frustration has been considerably less studied than the magnetic one , although recent works indicate that its effects may lead to rich physics , such as , robust superconductivity in strongly repulsive fermionic system @xcite and spontaneous time - reversal symmetry breakings @xcite , among others @xcite . in a seminal work , haerter and
shastry @xcite have found a @xmath2 antiferromagnetic nel order as the ground state of the @xmath3 triangular lattice hubbard model when the hole motion is frustrated ( @xmath4 ) , uncovering a new mechanism for itinerant magnetism . in this letter
, we further characterize this kinetic antiferromagnetism and we describe its microscopic origin , analyzing generic kinetically frustrated electronic models for which , in the limit of infinite coulomb repulsion and one hole doped away half - filling , the nagaoka s theorem is not valid . in particular
, we study the ground state of two hubbard models : one on the triangular lattice with a positive hopping term , and the other on the square lattice with positive second - neighbor hopping term . using the density matrix renormalization group ( dmrg ) @xcite
, we find in both cases that the ground state has antiferromagnetic order : @xmath2 nel order for the triangular lattice and the usual @xmath5 nel order for the square lattice .
surprisingly , we find that the local staggered magnetization becomes classical ( saturated ) in the thermodynamic limit
. this result can be thought as the almost - perfect antiferromagnetic counterpart of the nagaoka ferromagnetism ; the difference is that , as the local staggered magnetization does not commute with the @xmath6 invariant hubbard hamiltonian , classical antiferromagnetic states can not be the exact eigenvectors for finite lattices . based on a simple slave - fermion mean field @xcite
, we propose a mechanism responsible for the kinetic antiferromagnetism : if the hole were moving on a ferromagnetic background on these lattices , its kinetic energy would be frustrated .
however , when moving in certain antiferromagnetic background , the hole can release its kinetic energy frustration by , depending on the system , acquiring a non - trivial spin berry phase or having zero hopping amplitude along frustrating loops . as
the coulomb repulsion is infinite , no exchange interaction exist , being the stabilization of antiferromagnetism of pure kinetic origin .
_ hubbard model and dmrg .
_ we study the hubbard model , @xmath7 where we use the usual notation , and @xmath8 denotes pairs of neighbor sites connected by the hopping parameters @xmath9 . from the outset , we take @xmath10 we study the hubbard model on two lattices with frustrated kinetic hole energy : the triangular lattice with positive @xmath11 and the square lattice with nearest @xmath12- and positive next - nearest neighbor @xmath13 hopping terms . in the latter case ,
we choose @xmath14 as a generic point with kinetic frustration .
we take @xmath15 as the energy unit . to solve the hubbard model
we apply dmrg on ladders of dimension @xmath16 ( see fig . 1 in @xcite ) , with up to @xmath17 legs and @xmath18 rungs .
we choose clusters that are compatible with the antiferromagnetic orders found in this work .
this means that we take an even number of legs , @xmath19 , for both lattices , and an even ( multiple of 3 ) @xmath20 for the square ( triangular ) lattice .
we consider cylindrical boundary conditions with periodic wrapping in the rung direction , and open boundary conditions along the legs .
comparing results obtained with clusters of different number of legs , we find that the clusters with @xmath17 give a correct description of the two - dimensional systems .
a similar conclusion has been reached previously on square ladders @xcite .
we have kept the truncation error less than @xmath21 , assuring that errors of the dmrg are smaller than symbol sizes in each figure .
we have also consider the inclusion of a weak pinning magnetic field ( @xmath22 ) acting only on a single site , at the lower left - hand end of the clusters @xcite , or on one magnetic sublattice .
in previous works @xcite , the purpose of the small magnetic field has been to pin possible magnetic order , in order to reduce the computational efforts by computing the average value of the local spin instead of correlation functions , optimizing the truncation error . in our work ,
the inclusion of the weak magnetic field will allow us to highlight the classical character of the ground states .
_ static magnetic structure factor .
_ to detect the existence of magnetic order , we compute the static magnetic structure factor with dmrg for both lattices , @xmath23 where @xmath24 is the number of sites in the cluster , and @xmath25 run over all sites . in the inset of fig .
[ fig1](a ) we show an intensity plot of @xmath26 for a triangular cluster with @xmath17 legs and @xmath27 sites . for positive @xmath28
@xmath29 exhibits two sharp maxima at the momenta @xmath30 and @xmath31 corresponding to a three - sublattice @xmath2 nel order . as
these peaks diverge with increasing cluster size , the ground state exhibits a long - range order @xmath2 magnetic pattern .
this result has been obtained previously by haerter and shastry@xcite , diagonalizing an effective spin hamiltonian on smaller clusters , up to @xmath32 sites .
note that as we are working in the extremely correlated limit , where the exchange interaction driven by _ virtual _ kinetic processes vanishes , @xmath33 , the magnetic order can only have its origin in the hole motion .
on the other hand , as the triangular heisenberg model has the same @xmath2 nel order in its ground state @xcite , for finite @xmath0 ( @xmath34 ) and low doping , there is a synergy between _ real _ and _ virtual _ kinetic processes , that leads to the strengthening of the @xmath2 nel order respect to the half - filled case @xcite . for negative @xmath11 ( not shown in the figure )
the magnetic structure factor has a sharp peak at @xmath35 , while it vanishes for all other momenta .
this result correspond to a fully polarized ferromagnetic ground state , as predicted by the nagaoka s theorem . in the inset of fig .
[ fig1](b ) we show the magnetic structure factor for the @xmath3 hubbard model on square clusters with @xmath17 legs and @xmath36 sites , for @xmath14 . here
, there is a marked peak for the magnetic wave vector @xmath37 , corresponding to the usual two - sublattice nel order .
if we relax the infinite @xmath0 condition , the kinetic exchange interactions would favor a collinear antiferromagnetic order , characterized by the magnetic wave vector @xmath38 or @xmath39 , so there will be a competition between _ real _ and _ virtual _ kinetic processes , resulting in magnetic incommensuration and phase separation @xcite . in agreement with nagaoka s theorem , for @xmath40
the ground state is the saturated ferromagnet . for ( a ) @xmath3 triangular hubbard ( triangles ) and heisenberg ( circles )
models , without magnetic field ( open symbols ) and with a magnetic field @xmath22 applied to one sublattice ( solid symbols ) ; ( b ) @xmath3 square hubbard model with first- and second - neighbor hopping terms ( @xmath41 ) , without magnetic field ( open squares ) and with a magnetic field @xmath42 applied to one sublattice ( solid squares ) . dashed lines : classical local magnetization .
insets : intensity plot @xmath29 for the ( a ) triangular and ( b ) square models .
darker color indicates larger magnetic structure factor . @xmath43 and @xmath5 for the triangular and square lattice , respectively.,scaledwidth=46.0% ] _ local magnetization .
_ once computed the magnetic structure factor , we can get the order parameter for the antiferromagnetic order , the local staggered magnetization @xmath44 @xcite , where @xmath45 denotes the magnetic sublattices .
the cluster - size dependence of @xmath46 for the @xmath3 triangular hubbard model is shown in fig .
[ fig1](a ) ( open triangles ) , along with the local magnetization of the triangular heisenberg model ( open circles ) for comparison .
the dashed line indicates the classical local magnetization , @xmath47 corrected by the presence of the hole uniformly distributed ( as is confirmed by the dmrg calculations ) .
surprisingly , @xmath46 is very close to the classical value , even for small clusters , and it reaches this value in the thermodynamic limit ( @xmath48 ) . on the other hand , in the heisenberg case , strong zero - point quantum fluctuations lead to a drastic reduction of @xmath46 @xcite , and due to the quasi - one dimensional character of the clusters , @xmath46 extrapolates to 0 @xcite .
one possible reason for the classical character of the magnetic order is that the effective spin model , obtained after integrating the hole degree of freedom , contains effective long - range interactions ( see eq . 3 in ref @xcite ) that may favor the classical ordering , like in the lieb - mattis model @xcite . as we can see in fig .
[ fig1](a ) , @xmath46 for finite size clusters does not take _ exactly _ the classical value .
the hubbard hamiltonian is @xmath6 spin rotational invariant and it does not commute with the antiferromagnetic order parameter ; consequently , its ground state can not break this symmetry for finite systems .
instead , it is expected that the finite size ground state is a singlet , and that , only in the thermodynamic limit , there can be a spontaneous @xmath6 symmetry breaking driven by the collapse of many low lying states onto the ground state @xcite
. the existence of a tower of states for the @xmath3 triangular hubbard model has been confirmed in ref .
@xcite .
we argue that the small departure of the order parameter from the classical value is related with the singlet character of the finite size ground state , and not with zero - point quantum fluctuations that reduce the order parameter like in quantum antiferromagnets .
pictorially , the finite size ground state can be thought as a linear combination of several classical antiferromagnetic states lying in different planes . to strengthen this picture
, we apply a small uniform pinning magnetic field , @xmath49 in one sublattice only ( if we apply the magnetic field in only one site , the difference is quantitatively small , of only a few percent ) . fig .
[ fig1](a ) shows @xmath46 for the hubbard model with the magnetic field applied ( solid triangles ) and the same for the triangular heisenberg model ( solid circles ) .
it can be seen that in the hubbard model now @xmath46 becomes classical , because the magnetic field select one of the classical orders that compose the finite - size ground state . on the other hand , the magnetic field increases @xmath46 of the heisenberg model , but there remains strong zero - point quantum fluctuations .
fig .
[ fig1](b ) shows @xmath46 for the hubbard model on the square lattice , with ( solid squares ) and without ( open squares ) an applied uniform magnetic field in one sublattice .
the same behavior as in the triangular case is found : the local magnetization is close to the classical values when @xmath50 , being enough to apply a small @xmath22 to pin one classical magnetic ground state .
_ energy scale._for the triangular lattice , the extrapolated ground state energy is @xmath51 , in agreement with the value obtained in @xcite ( @xmath52 ) , while for the square lattice the extrapolated value is @xmath53 . in order to quantify the energy scale of the kinetic antiferromagnetism
, we match the effect of the hole motion to an effective nearest - neighbor antiferromagnetic heisenberg interaction , @xmath54@xcite , resulting @xmath55 , this is the energy difference _ per site _ between the fully - polarized ferromagnetic state and the antiferromagnetic ground states .
we have found that @xmath56 ( @xmath57 ) for the triangular ( square ) lattice , for large @xmath58 _ release of the kinetic frustration . _
now we trace back the origin of the kinetic antiferromagnetism by means of a comprehensive mean field approximation . to this end
, we use the slave - fermion schwinger - boson representation of the projected electronic degree of freedom in the @xmath59 model , the strong - coupling limit of the hubbard model . here
, we give a brief description of the mean field approach ( see supplemental material for the details @xcite ) . in this representation ,
the projected electronic operator is written as @xmath60 a composition of a schwinger boson @xmath61 , that account for the spin degrees of freedom , and a spinless slave fermion @xmath62 , that describes the charge sector .
this representation is replaced in the @xmath59 hamiltonian , resulting @xmath63 where we have defined the @xmath6 invariant operator @xmath64 related with ferromagnetic correlations between sites @xmath65 and @xmath66 @xcite , while @xmath67 describes the hole hopping amplitude .
@xmath68 is a bosonic term that represents the spin fluctuations due to the heisenberg term , and it vanishes when @xmath69 @xcite . after a mean field decoupling , we get @xmath70 where the hole kinetic energy dispersion takes the form @xmath71 ( @xmath72 are the relative position vectors of the sites connected by @xmath73 ) . in the one hole case , the ground state energy of the system corresponds to the bottom of @xmath74 .
this energy dispersion is tight - binding like , with the hopping terms @xmath73 renormalized by the ferromagnetic mean - field parameter @xmath75 @xmath76 .
the presence of the @xmath77 s parameters has two consequences : _ i _ )
on one hand , the hopping terms are renormalized as in the double - exchange mechanism @xcite , @xmath78 where @xmath79 is the angle between the spins separated by vector @xmath80 ; if the spins are antiparallel @xmath81 vanishes ; _ ii _ ) on the other hand , the renormalization can give rise to a non - trivial spin berry phase for non - collinear orders , encoded in the @xmath82 s signs and associated with the solid angle subtended by the spins on a closed loop @xcite .
when the system is kinetically frustrated , these two features of the hopping renormalizations act releasing the hole kinetic energy frustration as the hole moves through certain antiferromagnetic patterns .
now we describe how this release works in the triangular and square lattice cases . _
triangular lattice : _ for the ferromagnetic state all @xmath83 parameters are equal to @xmath84 , consequently the hole motion is frustrated for @xmath85 . on the other hand , in a @xmath2 nel order
the parameters @xmath86 become negative and the others remain positive since @xmath87 @xcite .
these negative @xmath77 s turn upside down the hole dispersion , releasing the kinetic frustration of the hole motion .
notice that the flux of the @xmath77 s parameters in a closed loop is the solid angle subtended by the magnetic order , and consequently it is associated with the spin - berry phase detected by the hole ( see supplemental material for details @xcite ) . _
square lattice : _ in this case , when @xmath88 the hole motion in a ferromagnetic state is frustrated .
however , in the @xmath5 nel order the vanishing of the effective first - neighbor hopping terms , due to their antiparallel spins @xcite , removes the frustrating loops , releasing the kinetic frustration .
we remark that , at the mean field level , the 120@xmath89 nel ( @xmath5 nel ) state is degenerate with the ferromagnetic one in the triangular case with @xmath85 ( square lattice with @xmath90 ) .
the reason for this degeneracy is that , although the hole motion through the antiferromagnetic states is not frustrated , there is a hole dispersion bandwidth reduction due to the hopping renormalizations @xcite .
so , strictly speaking , the mean field numerics does not show the stabilization of the kinetic antiferromagnetism over the nagaoka s state . however , after considering the combined effects of quantum interference and strong - correlation physics beyond the mean field approximation ( like our dmrg predictions ) , the actual kinetic antiferromagnetism emerges .
despite the mean field discrepancy , we strongly emphasize that the mean field approach allows to find one of the main ingredient of the kinetic antiferromagnetism , that is , the release of the kinetic frustration . using the insight we gained from the mean field approach
, we can predict the appearence of this novel kinetic antiferromagnetism phenomenon in other kinetically frustrated systems , like the anisotropic triangular lattice and the @xmath91 square lattice with @xmath92 preliminary dmrg results show the ubiquity of kinetic antiferromagnetism in these systems ( see supplemmental material @xcite ) .
furthermore , we remark that the kinetic antiferromagnetism mechanism is completely different to the exchange one , and in general the ground state selected by this itinerant mechanism does not have to be necessarily the classical ground state of the related heisenberg model . in this context ,
in particular , the @xmath3 kagom hubbard model with one hole doped and @xmath93 may be a promising candidate for the search of unconventional kinetic antiferromagnetism physics @xcite . finally , if we lift the condition of infinite @xmath94 it is possible to study the synergy between _ real _ and _ virtual _ kinetic processes , in order to highlight the crossover from the heisenberg regime , governed by the exchange interactions , to the kinetic antiferromagnetic one , governed by the kinetic energy . in fig .
[ fig2 ] we show the local magnetization of the ground state @xmath2 nel order of the triangular @xmath59 model predicted by the mean field approach and dmrg @xcite , as a function of @xmath95 for doping @xmath96 .
there is fairly good qualitative agreement between both methods , and for larger values of @xmath95 the order parameter is close to the heisenberg value calculated within each approach , @xmath97 @xcite and @xmath98 @xcite , while @xmath46 increases with decreasing @xmath95 , until it reaches the classical value for @xmath99 in both methods
. of the triangular @xmath59 model as a function of @xmath95 for doping @xmath100 the solid line corresponds to the mean field @xmath46 , the open circles to the dmrg results , the dashed line is to aid the eye .
the arrows indicate the @xmath33 and heisenberg @xmath46 limits.,scaledwidth=46.0% ] _ conclusions . _ using the density matrix renormalization group
, we find that classical antiferromagnetic ground states can be realized in extremely correlated electronic systems with frustrated kinetic energy .
in particular , we study the @xmath3 hubbard model , with one hole doped away half filling , on the positive @xmath11 triangular lattice , and on the square lattice with positive second neighbor hoppings .
we also propose a mechanism responsible for this kinetic antiferromagnetism , that is , the release of the kinetic energy frustration driven by , depending on the system , the spin - berry phase acquired by the hole while moving around an antiferromagnetic background , or the vanishing of the effective hopping amplitude along the frustrating loops .
this new mechanism for itinerant antiferromagnetism is quite ubiquitous for one hole doped away half filling in kinetically frustrated lattices @xcite , being also relevant in more general situations , like finite @xmath0 regime and low doping cases , as the mean - field results seem to indicate @xcite .
it is worth noticing that recent experiments @xcite were able to generate gauge fields which induce frustrated motion of ultracold bosons in triangular optical lattices , opening up the possibility to observe related kinetic antiferromagnetism phenomena .
we acknowledge useful discussions with c. d. batista and d. j. garca .
this work was partially supported by pip conicet grants 0160 and 0392 .
99 m. c. gutzwiller , phys .
rev .
lett .
* 10 * , 159 ( 1963 ) ; j. hubbard , proc .
r. soc .
a * 276 * , 238 ( 1963 ) .
p. fazekas , _ electron correlation and magnetism _ ( world scientific , singapore , 1999 ) . y. nagaoka , phys
. rev . * 147 * , 392 ( 1966 ) .
a. mielke , j. phys .
a * 24 * , lt23 ( 1991 ) ; h. tasaki , phys .
. lett . * 69 * , 1608 ( 1992 ) ; y. saiga and m. oshikawa , phys .
rev .
lett .
* 96 * , 036406 ( 2006 ) .
w. barford and j. h. kim , phys .
rev .
b * 43 * , 559 ( 1991 ) .
j. merino , b. j. powell , and r. h. mckenzie , phys .
rev .
b * 73 * , 235107 ( 2006 ) .
l. isaev , g. ortiz , and c. d. batista , phys .
rev .
lett .
* 105 * , 187002 ( 2010 ) .
o. tieleman , o. dutta , m. lewenstein , and a. eckardt , phys .
rev .
lett . * 110 * , 096405 ( 2013 ) .
y. f. wang , c.d .
gong , and z. d. wang , phys .
rev .
lett . * 100 * , 037202 ( 2008 ) .
z. p. yin , k. haule , and g. kotliar , nature materials * 10 * , 932 ( 2011 ) . j. o. haerter and b. s. shastry , phys
. rev .
lett . * 95 * , 087202 ( 2005 ) .
s. r. white and a. l. chernyshev , phys .
rev .
lett . * 99 * , 127004 ( 2007 ) .
s.r .
white and i. affleck , phys .
rev .
* 64*. 024411 ( 2001 ) . l. o. manuel and h. a. ceccatto , phys .
rev .
b * 61 * , 3470 ( 2000 ) .
a. weichselbaum and s. r. white , phys .
rev .
b * 84 * , 245130 ( 2011 ) .
l. liu , h. yao , e. berg , s. r. white , and s. a. kivelson , phys .
rev .
lett .
* 108 * , 126406 ( 2012 ) .
b. bernu , p. lecheminant , c. lhuillier , and l. pierre , phys .
rev .
b * 50 * , 10048 ( 1994 ) .
l. capriotti , a. e. trumper , and s. sorella , phys .
rev .
lett . * 82 * , 3899 ( 1999 ) . c. weber , a. laeuchli , f. mila , and t. giamarchi , phys .
rev .
b * 73 * , 014519 ( 2006 ) . c. n. sposetti , a. mezio , a. e. trumper , c. j. gazza , and l. o. manuel ( in preparation )
. the expressions are @xmath101 and @xmath102 for the triangular and square lattices , respectively .
e. lieb and d. mattis , j. math .
phys .
* 3 * , 749 ( 1962 ) .
j. o. haerter , m. r. peterson , and b. s. shastry , phys .
rev .
b * 74 * , 245118 ( 2006 ) .
w.f . brinkmann and t.m .
rice , phys .
rev .
b * 2 * , 1324 ( 1970 ) .
see supplemental material at http://link.aps.org/ ... h. a. ceccatto , c. j. gazza , and a. e. trumper , phys .
rev .
b * 47 * , 12329 ( 1993 ) . c. zener , phys . rev . * 82 * , 403 ( 1951 ) ; p.w .
anderson and h. hasegawa , phys . rev . *
100 * , 675 ( 1955 ) .
a. eckardt , p. hauke , p. soltan - panahi , c. becker , k. sengstock , and m. lewenstein , europhys .
lett . * 89 * , 10010 ( 2010 ) ; j. struck , c. lschlger , r.l .
targat , p. soltan - panahi , a. eckardt , m. lewenstein , p. windpassinger , and k. sengstock , science * 333 * , 996 ( 2011 ) . c. j. gazza and h. a. ceccatto , j. phys . : condens .
matter * 5 * , l135 ( 1993 ) . to compare similar results between the mean field @xmath46 computed for the two dimensional triangular lattice , and dmrg computations done in clusters with @xmath17 legs , we use the prescription given in ref .
@xcite , about choosing the optimal cluster shapes . in the case of 6 legs ,
this cluster is the @xmath103 ladder . |
the quantum information topics are the subject of active research @xcite .
the impressive progress have been reached in the quantum cryptography @xcite and study of quantum algorithms @xcite . due to an impact on security in quantum cryptography ,
the quantum cloning is still a significant topic . at the same time
, a cloning itself is hardly sufficient for an eavesdropping @xcite .
no - copying results have been established for pure states @xcite as well as for mixed states @xcite . in view of such evidences ,
the question arose how well quantum cloning machines could work . in effect , the basic importance of the no - cloning theorem is expressed much better in more detailed results , which also give explicit bounds on an amount of the noise .
after the seminal work by buek and hillery @xcite , many approaches to approximate quantum cloning have been developed . in view of existing reviews
@xcite , we cite only the literature that is directly connected to our results .
an approximate cloning of two prescribed pure states was first considered in ref .
this kind of cloning operation is usually referred to as _ state - dependent cloning _ @xcite . in general ,
various types of state - dependent cloners may be needed with respect to the question of interest @xcite .
errors inevitably occur already in a cloning of two nonorthogonal states @xcite .
how close to perfection can a cloning be ?
of course , any explicit answer must utilize some optimality criterion
. we will refer criterion used in ref .
@xcite to as the _ absolute error _ @xcite .
chefles and barnett @xcite derived the least upper bound on the global fidelity for cloning of two pure states with arbitrary prior probabilities . the quantum circuit that reaches this upper bound was also constructed @xcite .
the global fidelity of cloning of several equiprobable pure states was examined in ref .
@xcite .
although cloning problems were mostly analyzed with respect to the fidelity criteria , other measures of closeness of quantum states are relevant .
for example , the partial quantum cloning is easier to analyze with respect to the squared hilbert - schmidt distance @xcite .
one of criteria , _ relative error _
@xcite , has been shown to be useful within the b92 protocol emerged in ref .
@xcite . deriving bounds on the relative error
was based on the spherical triangle inequality@xcite and the notion of the angle @xcite sometimes called the _ bures length _ @xcite . using this new method ,
a cloning of two equiprobable mixed states was studied with respect to the global fidelity @xcite .
the results of ref .
@xcite were partially extended to mixed - state cloning @xcite . in a traditional approach
, the ancilla does not contain _ a priori _ information of state to be cloned just now .
a more general case is the scope of the stronger no - cloning theorem @xcite .
namely , a perfect cloning is achievable , if and only if the full information of the clone has already been provided in the ancilla state alone . in ref .
@xcite we examined a cloning of finite set of states when the ancilla contains a partial information of the input state .
so , the previous result of ref .
@xcite was extended to both the mixed states and _ a priori _ information . in this paper
, we study the relative error of cloning of several mixed states , having arbitrary prior probabilities . _
a priori _ information in the ancilla is also assumed . in section ii , the relative error criterion introduced in refs .
@xcite is extended to the general cloning scenario .
we derive the lower bounds on the relative error for cloning of two - state set ( see section iii ) and multi - state set ( see section iv ) . in section
v , the relative error is compared with other optimality criteria .
we also build the quantum circuit for cloning of two pure states ( see section vi ) .
this circuit reaches the lower bound on the relative error for arbitrary prior probabilities and _ a priori _ knowledge about the input .
section vii concludes the paper .
the main problem posed formally is this .
we have @xmath0 indistinguishable @xmath1-level systems that which are all prepared in the same state @xmath2 from the known set of density operators on the space @xmath3 .
these @xmath0 systems form the register @xmath4 .
its initial state is a density operator @xmath5 on the input hilbert space @xmath6 .
the prior probabilities @xmath7 of states @xmath8 obey the normalization condition @xmath9 .
we aim to get a larger number @xmath10 of copies of the given @xmath0 originals by means of the ancilla whose initial state is @xmath11 according to the input @xmath8 . here
we mean a system @xmath12 composed of extra register @xmath13 and environment @xmath14 .
the extra register @xmath13 contains @xmath15 additional @xmath1-level systems , each is to receive the clone of @xmath16 .
if we include an environment space then any deterministic physical operation may be expressed as a unitary evolution .
thus , the final state of two registers @xmath17 is the partial trace over environment space @xmath18 \ . \label{bd2}\ ] ] the output @xmath19 is a density operator on the output hilbert space @xmath20 .
the actual output @xmath19 must be compared with the ideal output @xmath21 .
many measures of distinguishability between mixed quantum states are based on the fidelity @xcite .
we shall employ the angles and the sine metric @xcite .
let @xmath22 denote a unique positive square root of @xmath23 .
the fidelity between the two density operators @xmath24 and @xmath25 is equal to @xmath26 @xcite . in terms of this measure ,
the angle @xmath27 $ ] between @xmath24 and @xmath25 is defined by the equality @xmath28 @xcite .
it is also referred to as the _ bures length _
@xcite , because of its close relation to the standard bures metric @xmath29 . due to the spherical triangle inequality @xcite , @xmath30
we introduce the _ sine distance _
@xcite between @xmath25 and @xmath24 as @xmath31 .
this metric on the space of quantum states has a close relation to the trace distance @xcite and enjoys the following @xcite .
for any povm measurement @xmath32 , there holds @xmath33 here @xmath34 is the probability of obtaining outcome @xmath35 , if the state right before measurement was @xmath25 . a more detailed characterization of such a kind can be posed via majorization relations @xcite .
we also have @xmath36 .
since the fidelity function can not decrease under any deterministic quantum operation @xcite , the last inequality can be extended to @xmath37 using the sine distance is reasonable approach due to the inequalities ( [ pr11 ] ) and ( [ pr14 ] ) . for brevity ,
let us denote @xmath38 when two inputs @xmath39 and @xmath40 are equiprobable , the relative error is defined by @xmath41 @xcite .
meaning @xmath42 , it can be rewritten as @xmath43 the right - hand side of eq .
( [ bd6 ] ) is quite relevant to the case of arbitrary prior probabilities . since the distance @xmath44 estimates the difference between two probability distributions ( see eq .
( [ pr11 ] ) ) , a reliable identification of original input via measurement over clones may be provided only when
@xmath45 so we see a reason for using a ratio of @xmath44 just to the half of @xmath46 .
in addition , this choice implies that the tight lower bound on relative error generally recovers the range @xmath47 $ ] .
we shall now extend the notion of relative error for the set @xmath48 with @xmath49 , when the number of different pairs is equal to @xmath50 .
the probability of taking the pair @xmath51 is equal to @xmath52 where @xmath53 .
we clearly have @xmath54 and @xmath55 for the set of @xmath56 equiprobable states .
to each pair @xmath57 assign the quantity @xmath58 which takes into account that , perhaps , @xmath59 .
it is natural to put the weighted average of the @xmath50 quantities ( [ aveid ] ) . *
definition 1 . * _ the relative error of @xmath60 cloning of the set @xmath48 is defined by _
let the prior probability be value of order @xmath62 for all the states except @xmath63 and @xmath64 .
that is , we take @xmath65 for @xmath66 , whence @xmath67 , @xmath68 for the rest pairs .
the expression ( [ bd7 ] ) for relative error is simply reduced to @xmath69 . in the same manner ,
we can find @xmath70 , when @xmath71 and probabilities @xmath65 except for the states @xmath72 solely .
we are interested in a nontrivial lower bound on the relative error ( [ bd7 ] ) .
our approach to obtaining the limits utilizes triangle inequalities @xcite .
following the method , we shall derive the angle relation from which bound on the relative error is simply obtained .
it is handy to introduce the angle @xmath73 $ ] as @xmath74 the laws of quantum theory impose some restrictions on acceptable values of angles @xmath75 , whence nontrivial bounds for different figures of cloning merit follow .
in the case of the two - state set @xmath76 , the initial state of ancilla is @xmath77 or @xmath78 according to the input which is @xmath79 or @xmath80 .
we further assume that @xmath81 and , by the multiplicativity of fidelity , @xmath82 .
the motivation is as follows .
if the inequality ( [ upsres ] ) is not satisfied then there are states sufficient for perfect cloning @xcite .
that is , there exist states @xmath77 and @xmath78 such that @xmath83 .
hence we can mention a trivial bound @xmath84 only .
so we presuppose that the inequality ( [ upsres ] ) is valid . as result , we have @xmath85 with no loos of generality , we assume that @xmath86 . * theorem 2 * _ the relative error @xmath87 of cloning of the set @xmath76 satisfies _ @xmath88 * proof * applying the inequality ( [ pr10 ] ) twice , we obtain @xmath89 recall that the fidelity function is multiplicative , preserved by unitary evolution and non - decreasing under the operation of partial trace @xcite .
so we obtain @xmath90 whence @xmath91 . combining this with eq .
( [ tilde1 ] ) provides @xmath92 consider the function @xmath93 to be minimized .
we want to minimize @xmath94 under the constraint ( [ star1 ] ) , @xmath95 and @xmath96 .
this task is solved in appendix a. by substitutions , we then have @xmath97 and further the statement of theorem 2 .
@xmath98 for equiprobable states , the bound ( [ theor3 ] ) is reduced to the lower bound deduced in ref .
@xcite . in terms of @xmath99 and @xmath100 ,
we rewrite ( [ upsres ] ) as @xmath101 . by @xmath102 and @xmath103 , the bound ( [ theor3 ] )
becomes @xmath104 at fixed @xmath105 and @xmath106 , the right - hand side of eq .
( [ theor33 ] ) is an increasing function of probability @xmath107 .
that is , it decreases as the prior probabilities differ .
this is analog of that the upper bound on the global fidelity increases in such a situation @xcite .
we are rather interested in dependence of the bound on @xmath106 .
this parameter marks a top amount of an _ a priori _ information , which can initially be laid in the ancilla .
the more a value of @xmath106 , the less this amount .
the angle @xmath108 is a decreasing function of @xmath106 . in the range ( [ alpdel ] ) ,
the lower bound by theorem 3 is a decreasing function of @xmath108 .
so the right - hand side of ( [ theor33 ] ) increases as the marker @xmath106 of additional information increases .
for @xmath109 the perfect cloning can be reached @xcite . in line with this fact
, we have @xmath110 and the vanishing bound on @xmath87 . on the contrary , in the usual cloning
there is no _ a priori _ information , i.e. @xmath111 and @xmath112 .
then the bound by theorem 2 reaches its maximum as a function of @xmath106 .
the above points reproduce the observations of ref .
@xcite in more general setting .
if @xmath113 at fixed @xmath114 then the right - hand side of the inequality ( [ theor33 ] ) goes to zero .
this is natural because infinite number @xmath0 of originals can provide almost perfect cloning . if @xmath115 at fixed @xmath0 then the right - hand side of ( [ theor33 ] ) recovers the value @xmath116 . in the standard cloning of equiprobable states ( @xmath112 , @xmath117 )
, this value can be arbitrarily close to 1 , since @xmath118 .
it is not insignificant that the value @xmath116 gives the minimal size of probability of inconclusive answer for unambiguous discrimination at @xmath117 .
namely , the success discrimination of the equiprobable pure states @xmath119 occurs with the optimal probability @xmath120 @xcite . the value @xmath116 is obtained for @xmath121 and @xmath122 .
note that the upper bound on the global fidelity in the limit @xmath115 at fixed @xmath0 goes to well - known helstrom bound @xcite .
it is the probability of correctly distinguishing between two pure states @xmath123 by the optimal strategy @xcite .
we now obtain a lower bound on the relative error of cloning of the set @xmath48 . as before ,
the prior probabilities are arbitrary and constrained only by the normalization condition .
like ( [ alpdel ] ) , we have the acceptable range @xmath124 according to theorem 2 , each term of sum in the right - hand side of ( [ bd7 ] ) obeys @xmath125 hence the desired bound is established as follows .
* theorem 3 * _ the relative error of @xmath60 cloning of the set @xmath48 satisfies _
@xmath126 as a straightforward extension , the bound ( [ theor4 ] ) succeeds many features of the bound ( [ theor3 ] ) .
if two probabilities , say , @xmath127 and @xmath128 are variable and the rest of parameters is fixed , then the bound ( [ theor4 ] ) decreases as these probabilities differ .
if some one probability is close to 1 and other probabilities are small , then the bound is close to zero .
this behavior is expected , because single known state can be cloned perfectly . for equal _ a priori _ probabilities @xmath54 , the bound by theorem 3 becomes @xmath129 it is natural that both the bounds given by ( [ theor4 ] ) and ( [ prieq ] ) decrease as @xmath130 increases .
indeed , the parameter @xmath130 characterizes an amount of prior information . if the upper limit of eq . ( [ alpdel2 ] ) is saturated for some pair @xmath57 then corresponding summands in the right - hand sides of eqs .
( [ theor4 ] ) and ( [ prieq ] ) vanish .
this is the case of potentially perfect cloning . on the whole ,
these conclusions on a role of _ a priori _ information in the ancilla add to the stronger no - cloning theorem .
a question is , whether the lower bounds ( [ theor3 ] ) and ( [ theor4 ] ) can be reached ? in general , it is not the case , though the bound by theorem 3 is least for two pure states .
the quantum circuit for optimal cloning will be built in the next section .
the subject matter changes for @xmath49 . from the viewpoint of minimization
the bound of theorem 3 is approximate . as reasons of appendix
a show , saturating the inequality ( [ rjk1 ] ) holds if and only if @xmath131 , @xmath132 for @xmath133 ( for @xmath134 the angles @xmath135 and @xmath136 should be swapped in the two equalities ) .
these two equalities per each of @xmath50 pairs totally give @xmath137 conditions . for saturating eq .
( [ theor4 ] ) , @xmath56 variables @xmath135 must satisfy all these @xmath137 conditions . except for some special cases , this is not possible .
thus , the presented limit is somewhat rough . more rigorous way may be as follows .
similar to ( [ star1 ] ) , we have arrived at the @xmath50 inequalities of a kind @xmath138 .
together with the @xmath56 conditions @xmath139 , these relations specify some simplex in @xmath56-dimensional real space . the relative error ( [ bd7 ] )
can be rewritten in the form @xmath140 where @xmath141 .
the task is to minimize the function ( [ bdel7 ] ) in the above simplex .
so we come across a difficult problem of nonlinear programming ( the simple case @xmath142 of this problem is considered in appendix a ) . for @xmath139 ,
the minimized function is concave .
so the problem of minimization is reduced to finding extremal points of the simplex .
if the values of parameters are prescribed , the wanted minimum can be found numerically . at the same time , it is complicated to obtain an explicit formula for general case .
but even if we should find it , we still would not have a complete solution to the problem of mixed - state cloning .
indeed , it is not necessary that bound given by such a formula be least .
so we have restricted our consideration to obtaining of the bound by theorem 3 .
rough though this bound is , it has straightforward form and allows to estimate how a merit of state - dependent cloning is limited .
we shall now expose the relative error in comparison with other optimality criteria . for the sake of simplicity
, we restrict to the @xmath60 cloning of two equiprobable pure states @xmath143 without _ a priori _ information in the ancilla .
how able to good cloning is the pair ?
this question is central to applications of quantum cloning . in principle
, we may assume both the deterministic cloning and probabilistic cloning @xcite .
a merit of deterministic cloning may be viewed with respect to the global fidelity , the absolute error and the relative error . for equiprobable inputs ,
the global fidelity is expressed by @xmath144 @xcite .
hillery and buek @xcite used the measure @xmath145 .
this measure will be referred to as _ absolute error _ @xcite .
the relative error is defined by eq .
( [ bd6 ] ) . in probabilistic cloning , the exact clone of an input successfully generated with the maximal probability @xcite @xmath146 where @xmath105 denote the overlap @xmath147 between states @xmath143 . as it is shown in refs .
@xcite , the maximum of the global fidelity is equal to @xmath148 according to ( [ theor33 ] ) , the minimum of the relative error is reduced to @xmath149 for @xmath112 and @xmath42 . for the absolute error
we have @xcite @xmath150 let us consider the two cases : ( i ) the states are @xmath143 are almost orthogonal , i.e. @xmath151 ; ( ii ) the states are @xmath143 are almost identical , i.e. @xmath152 with @xmath153 .
a behaviour of each of the criteria is shown in table 1 ( @xmath154 ) .
.an asymptotic behaviour of the four criteria .
[ cols="^,^,^",options="header " , ] as it is clear from the second column , for the case ( i ) all the measures endorse a good merit of both the deterministic and probabilistic cloning . in effect , the optimum of global fidelity is close to one , the optimum of absolute and relative error is close to zero .
the probability of success is close to one .
it is natural because orthogonal states can perfectly be cloned .
the principal distinction of the relative error is revealed in the case ( ii ) .
it seems offhand that two almost identical states can be cloned very well .
both the global fidelity and absolute error approve the conclusion ( @xmath155 and @xmath156 )
. it would be rash to accept this . in effect
, the optimal probability @xmath157 is generally not close to one .
the first term @xmath158 is almost one only if the number @xmath15 of actual clone is negligible in comparison with the number @xmath0 of originals . in line with this ,
the optimum of relative error is close to zero for @xmath159 .
but the probability is close to zero and the relative error is close to one when the number @xmath15 of actual clone is large .
we see that both the global fidelity and absolute error lose sight of the important aspect of deterministic cloning . even for the primary @xmath160 cloning
, we have @xmath161 and @xmath162 , that is both the probabilistic and deterministic strategies are restricted enough .
in contrast with the global fidelity and the absolute error , for the case ( ii ) a behaviour of relative error is crucially dependent on numbers @xmath0 and @xmath10 .
similar to the optimal probability of success , the criterion of relative error emphasizes that any cloning is not isolated stage in quantum information processing . as a rule ,
the outputs of cloning machine are subjects of further operations , say , a discrimination .
for example , in the cryptographic b92 scheme alice encodes the bits into two non - orthogonal pure states @xcite .
so bob can apply the unambiguous discrimination @xcite .
but the closer used states are to each other the larger number of discarded bits is in the total sequence . on the other hand , a sufficiently great closeness of the used states will prevent the eavesdropping . unlike both the global fidelity and absolute error ,
the notion of relative error allows to take such aspects into account .
we shall now build quantum circuits for the optimal relative - error cloning of two pure states @xmath143 with arbitrary prior probabilities @xmath163 ( @xmath86 ) . _ a priori _ information about actually input state is contained in the state of ancilla which is either @xmath164 or @xmath165 . without loss of generality ,
we take the product @xmath166 to be positive real .
these states are parametrized as @xmath167 , @xmath168 .
the overlap is @xmath169 with @xmath170 .
so , we have the register of @xmath171 qubits , where @xmath114 qubits are initially in the blank state @xmath172 , @xmath0 qubits are in the state to be cloned , and one qubit is ancillary . our aim is to transform these states according to the specification .
the strategy is an extension of the known one @xcite and uses the _ distinguishability transfer gate _ ( see appendix b ) .
first , the information about the input @xmath0 originals is transferred into one qubit .
we mark the ancillary qubit by 0 , the @xmath0 original qubits by @xmath173 , and the @xmath114 additional qubits by @xmath174 . the just left gate acts on the qubits @xmath175 and @xmath0 as @xmath176 where @xmath177 .
then an operation is applied to qubits @xmath178 and @xmath175 , and so on . in the first stage ,
the gate @xmath179 transfers the distinguishability from @xmath180th qubit to @xmath181th ( @xmath180 runs from @xmath0 to @xmath182 ) , i.e. @xmath183 where @xmath184 , @xmath185 . within the first stage ,
the state changes as @xmath186 where the gates @xmath187 are put from right to left with decreasing @xmath180 .
this part transfers a total distinguishability of the @xmath0 originals @xmath188 into the one - qubit state @xmath189 .
an example for @xmath190 cloning is shown on fig . 1 . cloning with the _ a priori _ information . for brevity ,
the gate @xmath191 in the left box and the gate @xmath192 the right box are both denoted as @xmath193 . ] for using an _ a priori _ information , we now include the _ turned _ gate @xmath194 .
this gate transfers distinguishability of ancilla s states @xmath195 to those of qubit 1 , namely @xmath196 after the action of gate @xmath194 , the ancilla contains no information about distinguishability .
all the distinguishability of inputs are now concentrated on two possible states @xmath197 of qubit 1 , where @xmath198 . now the scheme acts on the qubit 1 by the unitary operator @xmath199 specified as @xmath200 the values of angle @xmath201 and complex numbers @xmath202 and @xmath203 will be found below .
so the second stage results in the final state @xmath204 of the @xmath10 qubits . in fig .
1 , the gates @xmath194 and @xmath199 between the dash boxes perform the second stage . its structure is independent of numbers @xmath0 and @xmath10 . note that this stage and an _ a priori _ information are not considered in ref.@xcite .
we put two linear combinations of the ideal outputs @xmath205 and @xmath206 as @xmath207 the final stage of cloning is posed as @xmath208 .
let us continue the sequence @xmath209 with respect to the above recurrence , that is @xmath210 where @xmath211 $ ] .
hence we obtain @xmath212 . due to the property ( [ dp3012 ] ) of distinguishability transfer gate , we have @xmath213 in the third stage , the label @xmath214 in ( [ gdll ] ) runs from @xmath215 to @xmath216 .
so , the gate @xmath217 acts on the qubits 1 and 2 , the gate @xmath218 acts on the qubits 2 and 3 , and so on .
the total action is described by @xmath219 where the gates @xmath220 are put from right to left with increasing @xmath214 . in ( [ stag3 ] )
, the accumulated distinguishability is distributed among the @xmath10 qubits of interest . on fig .
1 , the four gates @xmath221 , @xmath222 , @xmath223 and @xmath224 of the third stage are grouped in the right dash box . using the linearity
, we see that @xmath208 too . due to @xmath225 , @xmath226 that is actually correct . specifying concrete values of @xmath202 and @xmath203 and herewith the single - qubit gate @xmath199 in eq .
( [ datet ] ) , we can optimize either the relative error or the global fidelity . in each case
, we superpose the @xmath227 onto the @xmath228 .
then after the second stage the @xmath10 qubits of interest lie in the states @xmath229 . for the optimality with respect to the relative error , we demand that @xmath230 , whence we get @xmath231 and @xmath232 , @xmath233 from ( [ tildd3 ] ) .
the angle between @xmath234 and @xmath235 is equal to @xmath236 , the angle between @xmath237 and @xmath238 is equal to @xmath239 . because unitary transformations preserve angles , the angle between @xmath240 and @xmath238 is equal to @xmath241 within the third stage , the state @xmath242 maps to @xmath243 . by definition , the value @xmath244 is angle between @xmath243 and @xmath245 . since @xmath246 and @xmath247 , we find the needed value @xmath248 .
thus , the inequality ( [ theor3 ] ) is saturated too , and the built scheme is really optimal with respect to the relative error .
note that @xmath249 and @xmath250 are found as @xmath251 and @xmath252 .
but the described geometrical picture is quite sufficient for all the purposes . in the same manner ,
the optimization of cloning with respect to the global fidelity would be considered . as result , the generalization of the deterministic cloner of ref .
@xcite to prior ancillary information can be obtained .
we have analyzed a new optimality criterion for the state - dependent cloning of several states with arbitrary prior probabilities and an ancillary information .
the notion of the relative error has been extended to the general cloning scenario .
the lower bounds on the relative error have been obtained for both the two - state and multi - state cases .
the attainability of the derived bounds has been discussed .
the quantum circuit for optimal cloning of two pure states with respect to the relative error has been built .
our approach is based on the simple geometrical description , which generally clarifies origins of a bound for one or another figure of merit . in principle , the described scheme allows to develop cloning circuit that is optimal with respect to any non - local figure of merit . the scenario with an _ a priori _ information in the ancilla was inspired by the stronger no - cloning theorem .
the obtained conclusions on a possible merit of the cloning contribute to this subject .
unequal prior probabilities of inputs are usual in communication systems .
the examination of mixed - state cloning is needed because all the real devices are inevitably exposed to noise .
analysis with respect to the relative error may have potential applications to the problem of eavesdropping in quantum cryptography .
let us consider the function @xmath253 , where positive @xmath254 and @xmath255 obey @xmath256 .
let @xmath257\,$ ] be a fixed parameter .
the range of variables is stated by conditions @xmath258 , @xmath259 and @xmath260 .
this domain @xmath261 is a square whose left - lower corner is cut off by line @xmath262 . * lemma 4 * _ the global minimum of the function @xmath263 in the domain @xmath261 is equal to _ @xmath264 .
* proof * inside of the domain @xmath261 , we have @xmath265 and @xmath266 . so the extreme values are reached on the boundary @xmath267 .
consider those segments that are parallel to either axis @xmath268 or axis @xmath269 .
the minimum value on these segments is equal to either @xmath270 or @xmath271 , i.e. @xmath272 . on the segment @xmath273
, we put @xmath274 and @xmath275 with @xmath276 $ ] , whence @xmath277 . by calculus
, we obtain the extreme value @xmath278 for @xmath276 $ ] .
this value is not less than both the @xmath279 and @xmath280 . @xmath98
by this operation , a distiguishability of the possible states of second qubit is translated to those of the first .
it is convenient to introduce a family of states @xmath281 with the inner product @xmath282 , where @xmath283 $ ] . as is well - known , one- and two - qubit gates are sufficient to implement universal computation . in the context of cloning , the writers of ref .
@xcite note that only one type of pair - wise interaction is needed .
the _ distiguishability transfer gate _ is described by @xcite @xmath284 where by the unitarity @xmath285 .
it follows from eqs .
( [ dp1230 ] ) and ( [ dp3012 ] ) that the operation @xmath286 is hermitian @xcite .
the action of distiguishability transfer gate on two - qubit register is shown on figure 2 .
the corresponding circuit of @xmath287 elements and one - qubit operations is given in ref . | the relative error of cloning of quantum states with arbitrary prior probabilities is considered .
it is assumed that the ancilla may contain some _ a priori _ information about the input state to be cloned .
the lower bound on the relative error for general cloning scenario is derived .
both the case of two - state set and case of multi - state set are analyzed in details .
the treated figure of merit is compared with other optimality criteria .
the quantum circuit for optimal cloning of a pair of pure states is constructed . |
recent years have seen a growing realization that the cores of rich galaxy clusters are complex and dynamic environments . in particular , it is becoming clear that the radio - loud active galactic nuclei ( agn ) often hosted by the cd galaxies in rich clusters can have a major influence on the hydrodynamics and thermodynamics of the core regions of the intracluster medium ( icm ) . as we discuss below , the rich datasets coming from the _ chandra x - ray observatory _ and _ xmm - newton _ now demand theoretical models that go beyond the simple picture of jet - blown `` bubbles '' rising in a static icm described by ideal hydrodynamics .
bulk icm motions ( including turbulence ) , magnetohydrodynamics ( mhd ) , thermal conductivity , and viscosity may all be relevant to data that are currently being taken . even before the launches of _ chandra _ and _ xmm - newton _ , _ einstein _ and _ rosat _ studies revealed prominent radio - galaxy / cluster interactions in three clusters ; perseus - a ( bhringer et al .
1993 ; heinz , reynolds & begelman 1998 ) , virgo - a ( feigelson et al .
1987 ; bringer et al . 1995 ) and cygnus - a ( carilli , perley & harris 1994 ; harris , carilli & perley 1994 ) .
_ rosat _ showed perseus - a and cygnus - a to possess intracluster medium ( icm ) cavities coincident with the prominent radio - lobes in these two sources , suggesting supersonic inflation of a bubble in the icm by the jetted agn ( clarke , harris & carilli 1997 ) .
virgo - a , on the other hand , showed a cooler ( and soft x - ray brighter ) region of icm associated with the outer eastern `` ear '' seen in low - frequency radio maps of virgo - a ( owen , eilek & kassim 2000 ) .
it was first suggested by bhringer et al .
( 1995 ) that this phenomenon might be caused by lower entropy gas from the cluster center being dragged upwards in the icm atmosphere by a buoyantly rising radio - lobe .
more recent observations by _ chandra _ and _ xmm - newton _ show that radio - galaxy induced icm substructure is surprisingly ubiquitous and complex .
the basic results described above , i.e. , the existence of icm cavities associated with active radio - lobes and the presence of cool material that appears to lie in the wake of old , buoyantly rising radio - lobes , have been confirmed in numerous systems ( e.g. , hydra - a ; mcnamara et al . 2000 , abell 2052 ; blanton et al .
2001 , virgo - a ; young , wilson & mundell 2002 , perseus - a ; fabian et al .
2000 , 2003a ) .
however , these new data have raised several mysteries and are increasingly at odds with simple models for radio - galaxy / icm interactions .
firstly , in any ideal hydrodynamic model , the cavities must be inflated supersonically or else they would be destroyed by rayleigh - taylor ( rt ) instabilities faster than they are inflated .
curiously , the strong and hot icm shocks that one expects to find around the active cavities are notably absent ; instead , many icm cavities are surrounded by icm shells that are _ cooler _ than the ambient icm
. we shall refer to this as the `` shock problem '' .
secondly , icm cavities that are not associated with any obvious radio - lobe ( `` ghost cavities '' ) have been discovered .
examples are found in the perseus cluster ( fabian et al .
2003a ) , abell 2597 ( mcnamara et al .
2001 ) , and abell 4059 ( heinz et al .
2002 , choi et al . 2004 ) . in some cases , ghost cavities are coincident with regions of low - frequency ( 74mhz ) radio emission supporting the hypothesis that they correspond to old radio - lobes from previous and now extinct agn outbursts ( fabian et al . 2002a ) .
interestingly , as we will explicitly demonstrate in this paper , ideal hydrodynamic models fail to reproduce the observed morphology of at least some ghost cavities .
finally , most clusters have been found to possess a `` temperature - floor '' in the sense that the radiative cooling of the icm appears not to proceed below temperatures of @xmath2 ( tamura et al .
2001 ; peterson et al . 2001 ) .
again , no clear explanation for this fact is provided by an ideal hydrodynamic model .
brggen & kaiser ( 2002 ) suggest that stirring of the icm core by a central radio galaxy may be responsible for the temperature floor ; however , such a scenario only postpones rather than prevents radiative cooling , and is hard to reconcile with the strong metallicity gradients observed in the cores of some clusters ( david et al .
2001 , matsushita et al . 2002 ; sanders & fabian 2002 ) .
this has led several authors ( narayan & medvedev 2001 ; fabian et al .
2002b ; voigt et al .
2002 ; kim & narayan 2003 ) to resurrect the notion that thermal conduction may be important in determining the thermodynamics of icm cores .
faced with the multiple failures of simple hydrodynamic models for cluster cores , we must carefully examine the other physical processes that might be relevant .
both the dynamical and thermodynamical effects of magnetic fields are largely unexplored and are the subject of on - going large - scale simulations .
however , the qualitative success of models including thermal conduction in solving the temperature floor problem begs a study of other transport processes and , in particular , the effects of shear viscosity on the dynamics of a radio - galaxy / icm interaction .
perseus - a and the core of the perseus cluster continues to be one of the best studied examples of a rich cluster core and a complex radio - galaxy / cluster interaction .
the most detailed x - ray investigation to date , based on a deep ( 200ks ) observation by _ chandra_/acis - s , has been presented by fabian et al .
( 2003a ) .
these authors report the discovery of wave - like disturbances in the icm on spatial scales of @xmath3 , approximately twice the spatial scales of the most obvious ghost cavity to the north - west of perseus - a .
they discuss a scenario in which viscous dissipation of these disturbances may act as a significant heat source for the icm core .
it is shown that , provided viscosity operates reasonably close to its ideal unmagnetized value , it is possible for viscous dissipation of radio - galaxy induced disturbances to balance radiative cooling of the icm .
this suggestion has been supported by recent simulation work by ruszkowski , brggen & begelman ( 2003 )
. circumstantial evidence for the presence of significant icm viscosity is also provided by an examination of the morphology of h@xmath1 filaments .
several of the filaments appear to trace well - defined arcs in the region below the ghost cavity ( fig .
[ fig : perseus ] ; also see fabian 2003b ) .
this argues against the presence of strong turbulence in the icm core , possibly resulting from the action of viscosity .
if we make the stronger presumption that the h@xmath1 filaments follow streamlines in the icm , the morphology of the filaments suggests the existence of a vortex ring within the icm just below the nw ghost cavity . with this background and motivation
, this paper presents hydrodynamic simulations of the buoyant evolution of an agn - blown cavity in a viscous icm . in order to bring clarity to the discussion of such a complex system
, this paper deals with the focused question of how icm viscosity effects the observed morphology and associated flow patterns of old ( ghost ) cavities .
detailed investigations of the effects of viscosity on the growth of active cavities and the thermodynamic state of the icm will be addresses in future work .
section 2 reviews the importance of viscosity in typical clusters , and touches upon the robustness of viscosity in the presence of magnetic fields .
section 3 describes the basic set - up of our simulations , as well as our results on the morphology and flow patterns .
section 4 discusses some of the limitations of this work , and possible implications of icm viscosity .
finally , conclusions are presented in section 5 .
before discussing simulations of viscous systems , we shall address in brief the general issue of viscosity in the icm .
initially suppose that the icm can be described as a thermal fully - ionized plasma which is unmagnetized .
the relevant coefficient of viscosity is given by braginskii ( 1958 ) and spitzer ( 1962 ) as @xmath4 where @xmath5 is the temperature of the plasma measured in kelvin and @xmath6 is a coulomb integral .
scaling to a temperature of @xmath7 and using @xmath8 gives @xmath9 .
it is customary to measure the importance of viscosity in a fluid of density @xmath10 through the reynolds number , @xmath11 , where @xmath12 and @xmath13 are characteristic velocities and length scales of the system and @xmath14 . of course
, in a complicated system such as a radio - galaxy / icm interaction , there is no unique velocity and length scale and so it is not possible to define a universal reynolds numbers that characterizes the system .
however , provided one uses some consistent choice for @xmath12 and @xmath13 , the reynolds number is still useful as a means to parameterize the relative importance of viscosity between different agn / icm systems
. it also allows a means of matching the viscosity imposed in simulations with that expected in real systems .
we choose the maximum dimension of the bubble as our characteristic length scale , and half of the adiabatic sound speed ( i.e. a typical buoyancy - induced rise velocity ) as our characteristic velocity . scaled to the nw ghost cavity of the perseus cluster ( with number density @xmath15 , and @xmath16 which gives @xmath17 ) , gives a reynolds number of @xmath18 we note that this reynolds number is almost an order of magnitude smaller than the fiducial value of @xmath19 quoted by robinson et al .
( 2003 ) , primarily due to our higher ( and more realistic ) fiducial temperature .
thus , restating the conclusion of fabian et al .
( 2003a ) , viscosity may be relevant to the evolution of an agn induced bubble in the icm .
the major uncertainty is the effect that magnetic fields have on the macroscopic viscosity .
the case of a uniform field is readily analyzed ( spitzer 1962 ) .
the proton gyro - radius corresponding to any non - negligible magnetic field is very small , leading to extremely efficient suppression of the local coefficient of viscosity perpendicular to the magnetic field ; for typical icm conditions , the perpendicular coefficient of viscosity is suppressed by the enormous factor of @xmath20 ( spitzer 1962 ) .
however , the effective macroscopic viscosity in the case of a realistic magnetic field configuration ( which is almost certainly tangled , and may be chaotic ) is an open question .
a similar issue has recently been addressed in the context of thermal conduction . in that case , the local thermal conductivity is also suppressed by a very large factor perpendicular to the field .
however , the exponential divergence of neighbouring field lines in a chaotic field structure results in an effective thermal conductivity , @xmath21 , that is suppressed below the unmagnetized value , @xmath22 by only a factor of @xmath23 ( depending on the spectrum of fluctuations in the field structure ; narayan & medvedev 2001 ) .
while the tensorial nature of the viscous stress tensor prevents a precise mapping of the two problems , similar arguments may apply and we might expect the effective coefficient of viscosity to be suppressed below the unmagnetized value by some factor ranging from @xmath24 to unity . clearly , the effective thermal conductivity and viscosity characterizing the icm is still very much an open theoretical question , due to uncertainties in both the basic physics of transport processes in hot plasmas as well as the magnetic field structure present in the icm . to make progress
we must _ assume _ that certain conditions exist , compute the consequences for radio - galaxy / icm interactions , and compare with the recent observations .
this is the motivation for the rest of this paper . while the evidence for non - negligible icm viscosity is still circumstantial ,
we show that the action of such a viscosity allows the morphology of ghost cavities ( in particular , the nw ghost cavity in perseus - a ) to be reproduced , and may be an important mechanism for stabilizing the icm core against radiative losses .
our goal is to study the evolution of a buoyant radio lobe in the icm of a galaxy cluster , including the effects of kinematic viscosity . following the approach of bruggen & kaiser ( 2001 ) ,
we simulate only the phase of the evolution after the radio - loud agn has inflated a low - density bubble which has expanded to achieve pressure equilibrium with the icm . at some point during this process
, the agn activity is assumed to cease .
thus , our simulation follows the evolution of a low - density bubble , initially in pressure equilibrium , placed in the central regions of an icm atmosphere . even though real radio lobes will have rapid and complex internal flows ( induced by the agn jet during their inflation )
, we shall assume that both the icm and the bubble interior are initially static .
we stress that this is an important simplification in our models and must be kept in mind when interpreting the results . for example
, our calculations will not be meaningful for addressing the issue of shocks and sound waves driven into the icm by the radio - galaxy , since these phenomena are almost certainly dominated by the jet - driven inflation phase of the bubble which we are not modeling .
furthermore , the precise growth rates of the rt and kelvin - helmholtz ( kh ) will be influenced by the internal motions within the bubble left over from its inflation phase .
however , our set - up allows a qualitative investigation of the hydrodynamics of buoyantly rising icm bubbles .
our detailed set - up is as follows .
the undisturbed icm atmosphere is given a density profile described by @xmath25^{-0.75}$ ] , where we choose units of mass and length such that @xmath26 and @xmath27 .
this atmosphere is assumed to be initially static and isothermal , with an adiabatic sound speed of @xmath28 .
the gravitational potential , @xmath29 , is assumed to be dominated by dark matter and , hence , is assumed to be fixed throughout the simulation .
this gravitational potential is determined by the condition that the initial icm atmosphere is in hydrostatic equilibrium , @xmath30 , where @xmath31 is the pressure . in this icm atmosphere
, we carve out a spherical bubble with density @xmath32 and radius @xmath33 .
this bubble is displaced from the center of the icm atmosphere by @xmath34 ( i.e. , the boundary of the bubble touches the center of the icm atmosphere ) .
the initial pressure of the bubble is set equal to the initial icm pressure at that radius ( i.e. , this is a hot , low - density bubble is in local pressure equilibrium with the icm ) .
we note that this is a very simplified form for the cluster density profile and gravitational potential .
however , we feel that a more sophisticated cluster profile ( e.g. , using a navarro , frenk & white [ 1997 ] profile , and including the potential of the cd galaxy ) would be unwarranted for the qualitative nature of the current investigation .
this initial state is evolved using the equations of 3-dimensional viscous hydrodynamics , @xmath35 where @xmath36 is the velocity field , @xmath37 is the internal energy density , and @xmath38 is the viscous stress tensor , @xmath39 note that we only include `` pdv '' and viscous dissipation terms in the energy equation , eqn . 4 . in particular , we neglect the radiative cooling which is unlikely to be relevant for the dynamical timescale processes modeled here .
these equations are solved using the zeus - mp mhd code ( stone & norman 1992a , 1992b ) , operating in cartesian co - ordinates @xmath40 .
zeus is a fixed - grid , time - explicit eulerian code which uses an artificial viscosity to handle shocks . when operated using van leer advection , as in this work , it is formally of second order spatial accuracy .
we model one half of the icm atmosphere using a simulation volume of a cubic box of length @xmath41 on a side .
the cluster center ( @xmath42 ) is placed in the center of one face of this cube , and reflecting boundary conditions are imposed on this face in order to account for the unmodelled half of the atmosphere .
outflow boundary conditions are imposed on all other faces .
the effects of viscosity are introduced into the basic zeus - mp code by adding explicit source terms to the momentum and energy equations ( eqns . 4 and 5 , respectively ) .
to ensure numerical stability , the time - step on which the equations were evolved was constrained to be @xmath43 where @xmath44 is the usual cfl timestep and @xmath45\ ] ] is the relevant viscous timestep . in this last expression , @xmath46 is the size of a grid cell in the @xmath47-th direction , and the minimization occurs over all grid cells .
on the basis of a standard stability analysis , we choose @xmath48 .
.basic information for the set of simulations presented in this paper .
[ cols="<,^,^,^,^",options="header " , ] in real systems , the interior of the icm cavities is filled with extremely tenuous and strongly magnetized relativistic plasma ; the coefficient of viscosity of this material is likely to be negligible .
however , our ( one - fluid ) simulations model these structures simply as bubbles of very hot gas ( with initial conditions of @xmath49 ) . if we were to include the temperature dependence of @xmath50 in our calculations ( eqn .
[ eq : mu ] ) , the coefficient of viscosity inside the bubble would be three orders of magnitude greater than in the ambient icm .
this is unphysical .
we address this problem by fixing @xmath50 to be constant in both space and time throughout a given simulation .
this choice suppresses the viscosity of the simulated hot gas inside the bubble .
we note that this method avoids the need to artificially truncate the viscosity within the bubble ( e.g. , see ruszkowski , brggen & begelman 2004 ) , leading to greater robustness and numerical stability .
we have , however , performed additional simulations in which we do truncate the viscosity within the bubble ( using a density threshold ) in order to assess the impact of this assumption .
comparing those simulations with the ones presented in this paper , we confirmed that the dynamics of the buoyant bubble considered here are not qualitatively effected by the constant @xmath50 assumption . using the same definition as section [ sec : viscosity ] ,
the reynolds number characterizing the hydrodynamics of the simulated bubble is @xmath51 where @xmath52 and @xmath13 are the characteristic velocity and size , respectively , of the bubble .
we perform a set of simulations to explore a range of reynolds numbers ; see table 1 for details of the runs performed .
initially we discuss the results from our zero - viscosity ( @xmath53 ) simulations .
these will serve as a crucial comparison for understanding the viscous simulations that we shall discuss next .
our qualitative simulation setup and results are very similar to those previously obtained by churazov et al .
( 2001 ) , brggen & kaiser ( 2001 , 2002 ) , brggen et al .
( 2002 ) and robinson et al .
of course , it is important to realize that even in this @xmath53 cases , the action of numerical diffusion keeps the effective reynolds number finite .
we determine the effective reynolds number of these `` zero - viscosity '' cases by performing additional simulations with small values of @xmath50 and visually comparing the smallest scale structures resulting from the @xmath53 run .
this exercise suggests that the effective reynolds numbers of our @xmath53 simulations are in the range 20005000 .
the initially static bubble starts accelerating due to buoyancy . as noted by several previous authors , rt instabilities
induce circulatory motions within the bubbles , that then induce `` secondary '' kh instabilities along the contact discontinuity between the low - density bubble and ambient icm .
these kh instabilities are primarily responsible for shredding the bubble within 23 time units ( i.e. @xmath54 sound crossing times of the bubble ; top - panels of fig .
[ fig : novisc ] ) . due to the shredding of the bubble
, one never observes a detached and flattened but otherwise intact structure such as we appear to see in the ghost cavity of perseus - a ( see bottom panels of fig .
[ fig : novisc ] ) . to further facilitate comparison with observations ,
we produce simulated x - ray surface brightness maps . in detail
, we set the local x - ray emissivity to be proportional to @xmath55 and integrate along lines of sight through the simulation volume in order to build up a 2-dimensional map of x - ray surface brightness .
for definiteness , we show surface brightness maps for the case in which the observer is viewing along the @xmath56-direction ( i.e. , the line joining the center of the icm atmosphere and the center of the initial bubble is perpendicular to the observers line of sight ) . from these maps , it can be seen that the observed cavity never has the appearance of the perseus - a ghost cavity ( fig .
[ fig : novisc ] ; lower panels ) .
in addition to morphological problems , we note that zero - viscosity models does not reproduce the h@xmath1-deduced flow pattern . at early times , while the buoyant bubble is still reasonably intact , we never see a circulatory flow pattern below the bubble . at late times , the buoyant bubble loses integrity and induces complex and disorganized motions in the icm .
having described the inviscid `` control '' case , we now proceed to discuss the effect of viscosity on the buoyant evolution of radio - lobes . as in the inviscid case ,
the evolution is driven by the joint action of buoyancy and secondary kh instabilities .
however , unlike the inviscid case where kt instabilities operate at the contact discontinuity on spatial scales down to the grid scale , viscosity suppresses the kh instability on small spatial scales .
this has a profound effect on the evolution of the bubble ; even a moderate amount of viscosity can prevent the shredding of the bubble , which can subsequently float out of the core being rather flattened but otherwise intact . as a specific example , fig .
[ fig : mu002 ] shows the @xmath57 ( run 3 ) .
this can be considered a model of the ghost cavities around perseus - a if the icm possesses a coefficient of viscosity of @xmath58 . as discussed in section [ sec : viscosity ] , this level of viscosity may be plausible even in the presence of tangled , chaotic magnetic fields .
it can be seen from the mid - plane density and velocity fields ( fig .
[ fig : mu002 ] ; upper panels ) that the evolution of the bubble is driven by buoyancy , with secondary kh instabilities largely unable to overcome the action of the viscosity .
as the bubble floats upwards , it flattens into a broad cap .
the surface brightness maps associated with run 3 ( fig . [ fig : mu002 ] ; lower panels ) show that one does , indeed , produce a detached and flattened cavity in the icm emission as observed in the perseus cluster .
figure [ fig : visc_range ] shows results for the full range of viscosity explored in this paper at three fixed times ( @xmath59 ) .
it can be seen that the formation of an flattened but intact buoyant bubble occurs in all of our viscous simulations . however , the timescale on which the evolution proceeds is a strong function of the viscosity .
for example , the @xmath60 density slice of run 2 ( @xmath61 ) is very similar to the @xmath62 slice of run 6 ( @xmath63 ) .
this fact may point to a solution of the `` shock problem '' noted in the introduction , an issue that we shall return to in section [ sec : shock ] .
viscosity also has important implications for the flow pattern in the disturbed icm . in principle
, the presence of viscosity can facilitate the development of large scale vortex rings in the trailing region beneath the rising bubble .
this phenomena is seen for our highest viscosity cases .
for the levels of viscosity that are probably relevant to the perseus cluster ( @xmath64 ) , this effect is not seen .
however , even for these levels of viscosity , our simulations show flow patterns that qualitatively match those inferred from the h@xmath1-filament geometry in perseus . in these cases ,
the flattened buoyant bubble undergoes a minor fragmentation due to the action of secondary kh instabilities . as a result of this fragmentation
, a small torus of radio plasma is left in the trailing region behind the main buoyant bubble .
the simulations show a strong icm circulation around this trailing torus , producing a streamlines that resemble the h@xmath1 filament geometry in perseus .
a prediction of this model is that sufficiently sensitive x - ray maps should reveal subtle depressions at the center of these vortices , and sufficiently sensitive and spatially resolved low - frequency radio maps should reveal a corresponding torus of aged radio plasma .
an important consequence of icm viscosity is that it provides an explicit mechanism by which the radio - galaxy induced disturbance can heat the icm . while a full treatment of viscous icm heating by radio - galaxies almost certainly requires following the jet - driven inflation of the bubbles ( e.g. , reynolds , heinz & begelman 2002 ) , as well as multiple epochs of activity ( ruzskowski , brggen & begelman 2004 ) , it is instructive to compute the viscous dissipation rate in our idealized simulation .
figure [ fig : dissipation ] shows the dissipation rate as a function of time for our canonical `` perseus - like '' model , the @xmath57 case .
the dissipation rate and the time coordinate are given in physical units assuming parameters relevant to the perseus cluster ( see caption of fig .
[ fig : dissipation ] ) .
the dissipation rate dispays two peaks .
the rather broad peak centered at about 90myr coincides with the secondary kh instability entering the strongly non - linear regime , and the subsequent `` folding '' of the flattened bubble .
the second and rather sharp peak corresponds to the venting of material out of the boundary of the simulation and , hence , should not be considered physical . in general , the dissipated power achieves the rather modest levels of @xmath65 .
however , this heat source continues to operate for a period of about 200myr , an order of magnitude more than the plausible recurrence timescale of the radio - galaxy activity ( fabian et al .
thus , the possibility of balancing the radiative losses with viscous dissipation from the combined effect of many bubbles remains open .
furthermore , it is likely that the dissipation of the fluid modes driven by the initial inflation of the bubble ( and hence not modelled here ) will deliver just as much energy , if not more , than the viscous dissipation during the buoyant phase .
our simulations must be viewed as a rather limited toy model for the buoyant evolution of radio - lobes and the formation of ghost cavities .
we have employed two major simplifications .
firstly , we have not modeled the jet - driven inflation of the bubble . real examples of these radio - lobes will possess complex , jet - driven flows with speeds much in excess of the icm sound speed , and will never resemble the static bubbles that we use for our initial conditions ( reynolds , heinz & begelman 2002 ) .
consequently , the kh instability will be important at all stages in the life of the bubble , and does not need to wait for the rt instability to first induce bulk flows . secondly , we have neglected the effects of magnetic field .
it is expected that fields within the radio plasma could readily achieve equipartition strengths and hence have an important effect on the dynamics of the rising bubble . in particular , magnetic fields will influence the development of both the rt and the kh instability at the contact discontinuity .
the effect of magnetic fields on the buoyant evolution of icm bubbles has recently been addressed in the context of inviscid hydrodynamics by brggen & kaiser ( 2001 ) , robinson et al .
( 2003 ) and de young ( 2003 ) .
their results suggest that magnetic fields can have a strongly stabilizing effect on buoyant bubbles , effectively quenching all of the instabilities that distort and eventually shred this structures .
however , both the symmetry of their calculation ( 2-d cartesian ) and their initial magnetic field configuration ( taken from cargill et al .
1996 ) makes the robinson et al .
( 2003 ) calculation most relevant to a buoyantly rising _ cylindrical flux tube with purely toroidal magnetic flux _ , a poor approximation to an agn blown bubble .
the ability of magnetic fields to stabilize or de - stabilize the contact discontinuity is a strong function of the magnetic configuration and field strength ( e.g. , see jun , norman & stone 1995 ) , both of which are open issues in the case of agn - blown bubbles .
clearly further simulations are required to address these two issues .
in fact , it would be most appropriate to treat these as coupled questions .
it is believed that the magnetic field in such bubbles originates from the magnetized agn jets .
thus , we must follow the jet - driven inflation of the bubble in order to produce a reasonable initial field configuration
. this will be the subject a future publication .
in addition to stabilizing buoyant radio - lobes against shredding , we have discussed how viscosity can dramatically slow the evolution of the bubble .
this suggests a natural solution to the `` shock problem '' noted in the introduction . under the action of viscosity
, the agn jets may inflate their icm bubbles subsonically before the buoyancy effects become important .
consequently , the icm can be displaced into a cavity - bounding shell without the need of a strong shock .
the thermal structure of this gas will be determined by the action of compression ( due to its displacement from the cavity ) , decompression ( due to the fact that this material will typically be lifted to a higher level in the icm atmosphere ) , viscous dissipation , and radiative cooling .
one can envisage a scenario in which the rather slow evolution of the radio - lobe in a viscous icm allows the shell of displaced icm to radiatively cool by an appreciable amount .
further simulations of the inflation stage of these bubbles are required to assess whether the rather cool cavity - bounding shells observed in per - a can be reproduced within the context of this model . given the strong temperature dependence of the coefficient of viscosity ( eqn .
1 ) , we might expect significant changes in the effective viscosity between the various icm structures defining the system ( i.e. , the ambient material , the cool rims and the shocks , if any ) .
thus , it will be important to move beyond the constant @xmath50 assumption in these next generation of simulations .
the true role of kinematic viscosity in icm / radio - galaxy interactions remains unclear , primarily due to the great uncertainty associated with magnetic suppression of spitzer - type transport processes in the plasma .
however , in stark contrast to the inviscid models , our models of bubble evolution in modestly viscous icm can reproduce both the morphology and inferred flow patterns seen around the best studied of the ghost icm cavities , the nw cavity in perseus - a .
the principal effect of viscosity is to stabilize the bubble against both rt and kh instabilities , thereby allowing it to remain intact as it floats upwards in the icm atmosphere .
the flattening of such bubbles as they rise is a natural explanation for the elongated morphology of the nw cavity in perseus - a .
the `` smoke - ring '' like flow pattern below the nw cavity inferred from the morphology of the h@xmath1 filaments can then arise due to minor fragmentation of the rising bubble ( leading to a trailing torus ) or genuine vortex shedding .
if the icm is indeed viscous at the level treated by our models , the subsequent slowing of the evolution timescales for icm / radio - galaxy interactions may be the solution to the `` shock problem '' , i.e. the lack of strong shocks bounding the icm cavities that are associated with active radio - lobes . in this scenario
, radio galaxies such as per - a would be rather less powerful than previously thought ( heinz , reynolds & begelman 1998 ) , inflating their associated icm bubbles subsonically .
if the evolution is sufficiently slow , radiative cooling might allow the icm shell surrounding the cavity to cool , again in agreement with chandra observations of perseus - a . ultimately , disentangling the effects of transport processes ( viscosity and thermal conduction ) and magnetic fields will require further simulation work coupled with detailed observations . on the theoretical side , it is important to initiate a new generation of simulations which incorporate full mhd _ as well as _ transport processes .
qualitatively new and rich dynamics becomes possible upon the inclusion of transport processes within mhd .
for example , balbus ( 2000 ) discovered the `` magnetothermal instability '' which occurs in any mhd atmosphere with a temperature profile that decreases with height provided there is thermal conduction along the magnetic field lines . in addition , the viscous transport coefficients will be anisotropic in the presence of a magnetic field , and this may lead to yet further complexity in the dynamics .
such instabilities could well be crucial for understanding the dynamics of radio - galaxy / icm interactions .
observationally , deep ( megasecond ) _ chandra _ observations of the nearest radio - galaxy / cluster interactions might allow detailed imaging spectroscopy of fluid instabilities associated with the radio - galaxy / icm interaction . furthermore , upon the launch of _ astro - e2 _ in 2005 , we will be able to directly probe turbulent velocities within the icm of nearby ( bright ) clusters through broadening of their x - ray emission lines . with the caveat that full mhd viscous simulations have yet to be done ,
the presence or absence of turbulence in the icm , especially in regions well separated from potential driving sources ( such as agn or merging subclusters ) could be an important indicator of the presence of viscosity .
we thank cole miller and eve ostriker for stimulating conversations conducted throughout the course of this work . we gratefully acknowledge support from the national science foundation under grant ast0205990 ( csr ) , the _ chandra _ cycle-5 theory & modelling program under grant tm4 - 5007x , and the royal society ( acf ) .
all simulations presented in this paper were performed on the beowulf cluster in the department of astronomy at the university of maryland , kindly administered by derek richardson .
balbus s.a . ,
2000 , apj , 534 , 420 blanton e.l . ,
sarazin c.l . ,
mcnamara b.r . , wise m.w . , 2001 ,
apj , 558 , l15 bhringer h. , nulsen p.e.j . , braun r. , fabian a.c . , bhringer h. , voges w. , fabian a.c . , edge a.c . ,
neumann d.m . , 1993 , mnras , 264 , l25 braginskii s.l . , 1958 ,
jetp , 6 , 358 brggen m. , kaiser c.r . , 2001 , mnras , 325 , 676 brggen m. , kaiser c.r . , 2002 , nat , 418 , 301 brggen m. , kaiser c.r . ,
churazov e. , enssline t.a . , 2002 , mnras , 331 , 545 cargill p.j . , chen j. , spicer d.s .
, zalesak s.t . , 1996 ,
jgr , 101 , 4855 carilli c.l . ,
perley r.a . ,
harris d.e .
, 1994 , mnras , 270 , 173 choi y.y . , reynolds c.s . , heinz s. , rosenberg j.l . , perlman e.s .
, yang j. , 2004 , apj , in press churazov e. , brggen m. , kaiser c.r . , bhinger h. , forman w. , 2001 , apj , 554 , 261 clarke d.a . , harris d.e . , carilli c.l . , 1997 , mnras , 284 , 981 david l. r. et al . , 2001 , apj , 557 , 546 de young d.s . , 2003 , mnras , 343 , 719 fabian a.c .
2000 , mnras , 318 , l65 fabian a.c .
, celotti a. , blundell k.m . ,
kassim n.e . ,
perley r.a . , 2002a , mnras , 331 , 369 fabian a.c . , voigt l.m
, morris r.g . , 2002b , mnras , 335 , l71 fabian a.c .
et al . , 2003a , mnras , 344 , l43 fabian a.c . , sanders j.s . , crawford c.s . , conselice c.j . , gallagher j.s . , wyse r.f.g . , 2003b , 344 , l48 feigelson e. , wood p.a.d . , schreier e.j . ,
harris d.e . ,
reid m.j . , 1987 , apj , 312 , 101 harris d.e . , carilli c.l .
, perley r.a . , 1994 , nat , 367 , 713 heinz s. , reynolds c.s . , begelman m.c . , 1998 , apj , 501 , 126 heinz s. , choi y.y .
, reynolds c.s . ,
begelman m.c . , 2001 ,
apj , 569 , l79 jun b. , norman m.l . ,
stone j.m . , 1995 , apj , 453 , 332 matsushita k. , belsole e. , finoguenov a. , bhringer h. , 2002 , a&a , 386 , 77 mcnamara b.r .
et al . , 2000 ,
apj , 534 , l135 mcnamara b.r .
et al . , 2001 ,
apj , 562 , l149 narayan r. , medvedev m.v . , 2001 ,
apj , 562 , l129 navarro j.f . ,
frenk c.s . ,
white s.d.m . , 1997 , apj , 490 , 493 owen f.n .
, eilek j.a . , kassim n.e . , 2000 , apj , 543 , 611 peterson j. a. et al . , 2001 , a&a , 365 , 104 reynolds c.s . , heinz s. , begelman m.c . , 2002 , mnras , 332 , 271 robinson k. et al . , 2004 ,
apj , in press ( astro - ph/0310517 ) ruszkowski m. , brggen , begelman m.c . , 2004 , apj , 611 , 158 spitzer l. , 1962 , the physics of fully ionized gases , interscience publishers , new york ( second edition ) .
stone j. m. , norman m. l. , 1992a , apjs , 80 , 791 stone j. m. , norman m. l. , 1992b , apjs , 80 , 81 tamura t. et al . , 2001 , a&a , 365 , l87 voigt l.m . ,
schmidt r.w .
, fabian a.c . , allen s.w .
, johnstone r.m . , 2002 , mnras , 335 , l7 young a.j .
, wilson a.s . ,
mundell c.g . , 2002 ,
apj , 579 , 560 | ideal hydrodynamic models of the intracluster medium ( icm ) in the core regions of galaxy clusters fail to explain both the observed temperature structure of this gas , and the observed morphology of radio - galaxy / icm interactions .
it has recently been suggested that , even in the presence of reasonable magnetic fields , thermal conduction in the icm may be crucial for reproducing the temperature floor seen in many systems .
if this is indeed correct , it raises the possibility that other transport processes may be important . with this motivation
, we present a numerical investigation of the buoyant evolution of agn - blown cavities in icm that has a non - negligible shear viscosity .
we use the zeus - mp code to follow the 3-dimensional evolution of an initially static , hot bubble in a @xmath0-model icm atmosphere with varying degrees of shear viscosity . with no explicit viscosity
, it is found that the combined action of rayleigh - taylor and kelvin - helmholtz instabilities rapidly shred the icm cavity and one does not reproduce the intact and detached `` ghost cavities '' observed in systems such as perseus - a . on the other hand , even a modest level of shear viscosity ( corresponding to approximately 25% of the spitzer value ) can be important in quenching the fluid instabilities and maintaining the integrity of the bubble .
in particular , we show that the morphology of the nw ghost cavity found in perseus - a can be reproduced , as can the flow pattern inferred from the morphology of h@xmath1 filaments .
finally , we discuss the possible relevance of icm viscosity to the fact that many of the active icm cavities ( i.e. , those currently associated with active radio - lobes ) are not bounded by strong shocks , the so - called `` shock problem '' .
= -1 cm [ firstpage ] cooling flows galaxies : jets hydrodynamics radio galaxies x - rays : galaxies : clusters |
the synchrotron emission from extragalactic radio sources is a function of both the local magnetic fields and the relativistic particle populations residing within .
these two components are important to the energy budget of such objects .
so , pinning down their relative contributions is crucial to understanding their overall behavior .
unfortunately , the optically - thin synchrotron emission alone can not be used to extract the individual particle and field components .
however , it has long been known that in principle radio synchrotron observations can be combined with x - ray observations of inverse - compton scattered cosmic microwave background photons ( hereafter ic/3k ) to extract information about particles and fields in emission regions ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
the advent of the chandra and xmm - newton observatories has made this kind of analysis possible for a large number of classes and objects .
x - ray emission has now been detected in the jets , hotspots , and lobes of numerous radio galaxies and attributed to a host of different physical processes ranging from synchrotron emission ( e.g. , @xcite ) to inverse compton scattered emission off of one or more of several photon fields . @xcite and @xcite reported the detection of x - ray emission in the lobes of 3c 219 and in 3c 215 and 3c 334 respectively , that possibly arises from inverse - compton scattering of ir photons from the quasar nucleus .
ic/3k lobe emission has been reported in several radio galaxies , including fornax a @xcite , centaurus b @xcite , abell85 0038 - 096 @xcite , 3c 295 @xcite , 3c 330 @xcite , and 3c 263 @xcite .
synchrotron self - compton ( ssc ) emission was reported in the hotspots of several powerful galaxies , including cygnus a @xcite , 3c 295 @xcite , 3c 273 @xcite , 3c 123 @xcite , 3c 207 @xcite , and 3c 263 and 3c 330 @xcite .
nonthermal x - rays of uncertain origin have recently been detected in the hotspots of 3c 280 and 3c 254 @xcite . when combined with radio observations ,
x - ray detections allow one to infer the magnetic field strength in an emitting region . yet , in practice the derived fields vary significantly , depending on the assumed mechanism for x - ray emission .
so , interpretation is often uncertain . indeed , closeness to the equipartition value is sometimes used as a validation criterion for a magnetic field estimated by any other means .
that seems unacceptable , since there is no convincing theoretical argument for equipartition between radio - emitting electrons and magnetic fields , and empirical evidence argues that not all sources are in equipartition ( e.g. , centaurus b @xcite , pks 0637 - 752 @xcite ) .
there are obvious complications with all of the field measures .
for example , deconvolution of the particle and field information also requires assumptions about the particle and field filling factors , as these values are impossible to extract from observations . for simplicity , a uniform magnetic field distributed through the emitting region
is customarily assumed . as explained below ,
caution is needed , particularly when assuming magnetic field uniformity .
it is important to be aware that the radio and x - ray emissions may be dominated by physically different regions in the source , and so one must be careful that the same particle population is sampled by the radio and x - ray observations .
this paper is based on the expectation that synthetic observations of numerically simulated radio galaxies may help us understand these issues in a way that makes minimal use of convenient , simplifying mathematical assumptions , while retaining the benefit of complete knowledge of the actual physical conditions being observed .
@xcite ( hereafter tjr01 ) recently carried out three dimensional time dependent mhd simulations of radio - jet flows that included nonthermal relativistic electron transport in space and momentum , enabling them to create the first synthetic radio observations from nonthermal electron distributions that were consistently evolved within the plasma flows .
those synthetic observations were used in conjunction with a detailed dynamical analysis to study how the dynamical and nonthermal particle transport processes lead to observable radio synchrotron surface brightness and spectral index patterns .
preliminary results of the synthetic observations were also reported in @xcite , @xcite , and @xcite . here
we look anew at the simulation data presented in tjr01 , shifting our attention from a dynamical analysis to the extraction of physical properties of the simulated objects using standard radio and x - ray observational analyses and then comparing results to the actual physical properties of the simulated objects . through this effort
we aim for new insights into how well standard analysis assumptions and techniques work to capture the true physical nature of a source .
in addition , we attempt to identify crucial issues for successful extraction of physical properties of a given object .
we emphasize that our purpose here is not to simulate specific real world radio galaxies .
neither do we intend to reevaluate the observations and analyses of particular sources .
additional details and discussion can be found in @xcite .
the paper is structured as follows .
[ bic : meth ] provides a brief exposition of our numerical methods , a review of the model parameters from tjr01 , and an overview of our synthetic observation techniques .
these techniques are then applied to the models in
[ bic : rad - x]-[bic : bme ] . in
[ bic : rad - x ] we look at what can be learned from the correlations between x - ray and radio surface brightness . [ bic : bic ] is devoted to an examination of the magnetic field inferred from synchrotron and inverse - compton surface brightness , @xmath0 and the minimum - energy field , @xmath1 . combined use of @xmath0 with the minimum - energy field @xmath1 as a possible tool to estimate local nonthermal energy content is discussed in
[ bic : bme ] .
the key findings are then summarized in [ bic : conc ] .
our numerical methods were detailed in @xcite and tjr01 . in short ,
we follow the bulk flow through a 3d tvd eulerian ideal mhd scheme and evolve a passive relativistic electron population on the eulerian grid through the standard particle kinetic equation .
our mhd code is based on an extension of the `` total variation diminishing '' scheme @xcite , as detailed in @xcite and @xcite .
the code preserves @xmath2 at each time step using a constrained transport scheme @xcite .
we include a passive `` mass fraction '' or `` color tracer '' , @xmath3 , to distinguish material entering the grid through the jet orifice ( @xmath4 ) from ambient plasma ( @xmath5 ) .
our electron transport scheme takes practical advantage of the mismatch between bulk flow and diffusive transport scales for gev electrons of relevance to radio and x - ray emissions within radio galaxies . in short , the lengths and times appropriate to the dynamics are orders of magnitude larger than those for electrons at energies relevant to radio synchrotron radiation .
the electron momentum ( energy ) distribution , @xmath6 , is then sufficiently broad that it can be adequately represented by a piecewise power - law form within a few broad bins in momentum space . at shocks ,
rapid diffusive acceleration for @xmath7 gev electrons ensures that they will emerge `` instantaneously '' from shocks with power - law momentum distributions .
subsequent , downstream cooling can be treated in a straightforward way .
we therefore divide the momentum domain into a small number @xmath8 of logarithmically spaced bins , and estimate particle fluxes across momentum bin boundaries by representing @xmath9 within bins , where @xmath10 varies in a regular way @xcite . in the test - particle limit for diffusive shock acceleration used here ,
electrons emerge from shocks with a power - law spectral index @xmath11 , where @xmath12 is the shock compression ratio .
the magnetic field is sufficiently weak in these simulations that all strong shocks behave dynamically essentially as hydrodynamic shocks . in accord with current understandings of collisionless shocks ( e.g. , @xcite )
we also inject and accelerate electrons from the thermal plasma passing through shocks .
injection physics at shocks is still poorly understood , so we simply assume that a small fraction , @xmath13 , of the thermal electron flux through a shock becomes extended via diffusive shock acceleration into the aforementioned power - law beginning at momenta just above the postshock electron thermal values . here
we give a short rundown of the physical parameters of the three simulations first introduced and analyzed in tjr01 .
the models are summarized in table [ summary.t ] , which is reproduced from tjr01 .
the simulated mhd flows of tjr01 are dynamically identical .
each jet entered the grid with a simple `` top hat '' velocity profile with @xmath14 , where @xmath15 is the sound speed in the uniform ambient medium .
the jets entered the grid at @xmath16 in gas pressure balance with the ambient medium and with a density contrast @xmath17 , giving a jet - based mach number @xmath18 .
the initial jet core radius was @xmath19 zones , while the entire @xmath20 uniformly zoned grid is @xmath21 . in units of initial jet radius ( @xmath22 ) and ambient sound speed ( @xmath23 , with @xmath24 ) the simulations stopped at @xmath25 time units when the bow shock reached the boundary at @xmath26 .
in physical units , the jet radius @xmath27 kpc and the inflow velocity @xmath28 .
this leads to a physical time unit @xmath29 years and a computational grid length @xmath30 kpc .
open boundary conditions were used everywhere except at the jet orifice .
the initial axial and background magnetic field ( @xmath31 ) ( also termed `` fiducial '' below ) exerted a magnetic pressure 1% of the ambient gas pressure ( @xmath32 ) .
in addition to the axial component , @xmath33 , the in - flowing jet also carried a toroidal magnetic field component derived from a uniform axial current and a return current on the jet surface ; i.e. , @xmath34 for @xmath35 . to break cylindrical symmetry , we added a modest wobble to the in - flowing jet velocity ; that is , it was slowly precessed on a cone of opening angle @xmath36 with five periods during the run .
tjr01 presented three idealized examples of electron transport , designed to isolate individual transport behaviors .
those models are also summarized in table [ summary.t ] . here , briefly , are other vital details .
electrons were transported explicitly over the momentum range @xmath37 with @xmath38 and @xmath39 ( with @xmath40 in units of @xmath41 ) for all models .
below @xmath42 the distribution function @xmath6 was continued as a power - law . at @xmath43
the gradient in the distribution function slope , @xmath44 was continued .
eight momentum bins ( @xmath45 ) were used for each simulation .
since the simulated nonthermal cosmic - ray electrons were passive , all results can be scaled for other choices of @xmath42 ( @xmath46 is fixed ) .
all three transport models included adiabatic cooling and diffusive shock acceleration , although second - order fermi acceleration and coulomb losses were neglected . in each model ,
the jet nonthermal electron population entered with a momentum index @xmath47 , representing a synchrotron spectral index , @xmath48 , where @xmath49 .
as listed in table [ summary.t ] we excluded local shock electron injection in models * 1 * ( hereafter called the `` control model '' ) and * 3 * ( hereafter called the `` cooling model '' ) , setting @xmath50 . those models isolated evolutionary behaviors of pre - existing electron populations in the jet
* 2 * the injection parameter @xmath51 , so we label model * 2 * the `` injection model '' .
details of these @xmath13 choices are given in @xcite and tjr01 .
we note that since the electron population in the simulation was passive , the injection model results could be simply rescaled for an alternate @xmath13 .
the in - flowing jet nonthermal electron population in this model was made much smaller than the other two models , so that local enhancements of the electron populations at shocks could be isolated .
the control model and the injection model are `` adiabatic '' in the sense that electrons experience negligible synchrotron aging .
that feature eliminates convex spectral curvature , although some concave curvature can result from spatial mixing of dissimilar power - law populations .
model * 3 * , on the other hand , includes significant radiative aging from synchrotron and inverse - compton processes , thus its cooling model label . in order to parameterize radiative particle aging , we defined a characteristic synchrotron cooling time , @xmath52 , for electrons with @xmath53 in the fiducial magnetic field . for the adiabatic models * 1 * and * 2 * , we set @xmath54 ( compared to @xmath25 ) by setting @xmath55 g and ignored inverse - compton losses , as well , to ensure negligible radiative aging for electrons of interest . for model * 3 * we made @xmath56 by setting @xmath57 g .
inverse - compton losses from the cmb were taken into account by including a term with @xmath58 g corresponding to the current epoch . once again
, the physical unit for @xmath59 in these simulations is approximately 10 myr .
our synthetic observation technique is straightforward .
we combine vector magnetic field and nonthermal electron distribution data from our simulations to calculate self - consistent radio and nonthermal x - ray volume emissivities in every zone of the computational grid . in order to simplify the analysis we restrict emissions to zones dominated by bulk plasma originating at the jet orifice , by requiring @xmath60 .
a raytracing procedure integrates along lines of sight in the optically thin limit to project emissions from the simulated objects on the sky for an arbitrary orientation .
we write the output data in fits format , and analyze it using conventional observational packages ( miriad and karma @xcite ) .
since the simulations are high resolution , the synthetic maps produced are of much higher resolution and dynamic range than typical real observations . to see the influence of resolution on our analysis we also have convolved the synthesized brightness distributions with circular gaussian beams using the miriad task convol to several lower resolutions . for convenience
, we placed all the objects at a fixed luminosity distance of 100 mpc , although the choice has no influence on our conclusions . for this distance ,
the nominal unconvolved resolution is 0.28 arcseconds , and the projected jet length is about 110 arcseconds .
we present results for convolved resolutions of 3.0 and 22.0 arcseconds as well , chosen to correspond to roughly 37 and 5 telescope beams along the jet , respectively .
these choices , while arbitrary , match qualitatively what is commonly achieved in many real source observations .
all synthetic observation images are set to 512 @xmath61 512 pixels ( not every pixel contains a nonzero brightness value , however ) . as in tjr01
we confine our discussion to one representative point in time ; namely , @xmath62 in simulation units or about @xmath63 yr . all synthetic observations in this chapter
refer to the same source orientation on the sky as that used in tjr01 , in which the jet axis is about 45 degrees from the plane of the sky .
our conclusions are unaffected by these choices .
the bulk flow is nonrelativistic , so no doppler corrections have been applied . in each spatial zone
we compute a synchrotron emissivity based on the local vector magnetic field , * b * , and nonthermal electron distribution , @xmath6 , as evolved by our transport scheme .
as given by @xcite , the emissivity is @xmath64 the spectral index @xmath65 is related to the local electron momentum index @xmath66 via @xmath67 , @xmath68 , where @xmath69 projects the local field onto the sky , and @xmath70 is an order - unity dimensionless constant , defined in @xcite . for a selected
observing frequency , @xmath71 , the distribution , @xmath72 , and the index , @xmath66 , are determined for each point on the grid by establishing the relevant electron momentum from the relation @xmath73^{(1/2)}$ ] , with @xmath74 in units @xmath41 .
we note for the magnetic fields in our simulated objects and radio frequencies in the ghz band that typical @xmath75 .
we compute an x - ray emissivity , @xmath76 , including inverse - compton contributions from the cmb ( hereafter ic/3k ) and synchrotron self - compton ( ssc ) ; viz , @xmath77 .
we ignore inverse - compton scattered agn photons , as well as x - ray synchrotron emission . although we include ssc emission , since its analysis does provide some useful insights , we mention that it is typically several orders of magnitude less intense in our simulated objects than ic/3k emission
. the x - ray ic/3k emissivity , @xmath78 , can be simply expressed at a selected x - ray frequency , @xmath79 , in terms of the synchrotron emissivity , @xmath80 , in equation ( [ synch.e ] ) extrapolated to @xmath79 ; namely @xcite , @xmath81 where @xmath82 and @xmath83 are the energy density and characteristic frequency , respectively , at the current epoch of the cmb , while @xmath84 is another order - unit constant that can be obtained from @xcite .
we note in equation ( [ cmb ] ) that @xmath80 is normalized and @xmath65 is determined at @xmath85 . for @xmath86kev ,
considered below , @xmath87 , which is substantially less than momenta responsible for the ghz synchrotron emission .
the ssc emissivity , @xmath88 , depends upon the synchrotron intensity distribution incident upon each zone . to keep the calculation manageable
, we adopt the common approximation of an isotropic incident intensity , so that @xmath88 can be expressed in terms of the omnidirectional incident flux , @xmath89 . except
when the radiation field is dominated by a very intense anisotropic local source , this approximation should be good , so adequate for our present purposes .
then , using a convenient expression in terms of the ic/3k x - ray emissivity from the same electron population we have @xcite , @xmath90 j_{3k}(\nu_x)~. \label{bic : jssc.e}\ ] ] here @xmath91 is the synchrotron spectral index at the low - frequency synchrotron cutoff @xmath92 , and @xmath65 is the spectral index determined for equation ( [ cmb ] ) .
the constant @xmath93 is order - unity and obtainable from @xcite .
we note that since the dominant synchrotron emission occurs at much lower frequencies than @xmath94 , electrons responsible for ssc emission are generally of higher energy than those producing the ic/3k emission .
we use a simple and fast fft - based scheme to estimate the omnidirectional synchrotron flux @xmath95 . details of the calculation can be found in appendix [ app : ssc ] .
we begin with a brief , qualitative comparison between the synthetic radio and x - ray brightness distributions .
radio surface brightness maps for the three models were shown in tjr01 . here , figures [ xray - m8.f ] through [ xray - m7.f ] show the corresponding ic/3k and ssc x - ray surface brightness images . from the images it is apparent that dominant dynamical features such as the jet and the radio hotspots are visible in both bands , but also that many details differ .
for instance , x - ray brightness distributions are smoother , with much less dynamic range than the radio distributions .
this simply reflects the fact that the inverse - compton brightness represents only the column density of relativistic electrons over a narrow range of electron energies .
the synchrotron brightness also depends strongly on the ( intermittent ) magnetic field distribution . in [ bic : bic ] and [ bic : bme ] we will combine the synthetic radio and x - ray data to explore their effectiveness for inferring source magnetic field properties . first , however , we demonstrate that some basic dynamical relationships connecting particles and fields can be extracted from correlations between the various intensity distributions , which are shown in figures [ x1vi.f ] , [ x2vi.f ] and [ x2vx1.f ] .
as we examine our three simulated objects , we should keep in mind that the magnetic field structures are identical in all three models , except for magnitude ; only the nonthermal electron distributions differ .
also , we emphasize , once again , that our goal is to examine ways of extracting reliable information about the objects we have in hand ; that is , how we extract from observations the physical properties that are actually present . because we know their properties exactly , the simulated objects are uniquely valuable testing grounds , independent of how closely those detailed properties match astrophysical objects
. figures [ x1vi.f ] and [ x2vi.f ] are scatter plots of 7.5 kev ic/3k and ssc intensities versus 2.9 ghz radio intensity , respectively , for the three electron transport models ( table [ summary.t ] ) .
similar distributions would be obtained for other observing frequency choices .
every other nonzero pixel is represented , for just under @xmath96 points in each plot . in each panel
the x - ray and radio brightness distributions are broadly correlated .
it is possible to extract from these trends insights about the dynamical behaviors of the source magnetic fields , as we illustrate .
we emphasize , however , that at a given radio ( x - ray ) brightness , the x - ray ( radio ) intensities generally spread over more than an order of magnitude , so it would be unrealistic to try to predict the inverse - compton brightness distribution from the radio distribution or vice versa .
this is especially so in the lobe structure , where the spread in synchrotron brightness is enormous , due to the wide range of magnetic field strengths .
comparison between figures [ x2vi.f ] and [ x1vi.f ] shows that the ssc brightness distributions of our simulated objects are generally several orders of magnitude less than the ic/3k intensities .
the difference is more than 10 orders of magnitude for both the control model and injection model , where the magnetic field magnitude was set very small to reduce synchrotron energy losses .
even for the cooling model , with its stronger field , the ssc intensities are mostly three to four orders of magnitude less than the ic/3k intensities .
this behavior comes from the fact that the synchrotron omnidirectional flux is generally much smaller than the cmb omnidirectional flux in our simulated objects .
that mirrors the situation in real radio galaxies , as well , since only the most intense x - ray hotspots have been found to be ssc dominated ( e.g. , * ? ? ?
* ) . despite this
, it is instructive to include the ssc brightness distributions in our analysis , since they reveal some useful insights .
the ic/3k vs radio trends in figure [ x1vi.f ] are well described for all three models by the proportionality @xmath97 , with @xmath98 . to understand this ,
suppose the local magnetic field scales with the electron density as @xmath99 .
since @xmath100 and @xmath101 , where @xmath102 is the path length , we have @xmath103 . on the other hand , by standard arguments @xmath104 for a disordered field in compression - dominated flows .
assuming @xmath105 , this gives @xmath106 .
for @xmath107 , typical of the synthetic radio sources , this would give @xmath108 .
alternatively , at perpendicular shocks one expects @xmath109 , resulting in @xmath110 , or @xmath111 .
thus , the observed trends are consistent with field evolution dominated by these dynamical processes . on the other hand , field evolution controlled by stretching of flux tubes would satisfy @xmath112 , where @xmath113 is the length of the flux tube ( e.g. , * ? ? ?
there is no explicit interdependence between density and magnetic field .
we know that the magnetic fields in the simulated object are filamented , so flux tube stretching certainly takes place .
the above exercise brings out the fact that the global field evolution in the simulated object is dominated by compression , however .
this matches our conclusions about the simulated field behaviors in tjr01 . sheared field evolution , while clearly involved and locally important , predominantly adds scatter to the correlations in figure [ x1vi.f ] . with the notable exception of the dominant hotspot in the injection model , the ssc and radio intensity distributions follow similar correlations as the ic/3k - radio correlation ; namely , @xmath114 , with @xmath115 . from this
match between ssc and ic/3k trends one might expect that @xmath116 , and this is nicely confirmed for the control model in figure [ x2vx1.f ] .
this proportionality is also evident in the other two models , excepting the injection model hotspot , although there is a lot more scatter .
for the general distribution we note from equation [ bic : jssc.e ] that one can write @xmath117 .
the observed correlations then imply , especially for the control model , that @xmath118 within the radio lobes .
that represents the fact that the average omnidirectional flux for the source is a good approximation to the local flux at any point in the source .
this follows if the omnidirectional synchrotron flux at each point is dominated by the collective contributions of distant emissions within the source . in that case
the average synchrotron emissivity becomes an effective value to use in estimating the ssc emission from the object .
the situation is much different in the dominant injection model hotspot .
the ssc / radio intensity correlation is described by @xmath119 , whereas the ssc and ic/3k intensities follow @xmath120 . on the other hand ,
the correlation between ic/3k and radio surface brightness is similar to what is seen elsewhere in the source and in the other two models .
previously , we interpreted the ic/3k vs radio correlation as evidence for magnetic fields evolving according to the scaling @xmath121 , which , therefore , seems applicable in the hotspot , as well .
it is the ssc behavior that is different in the injection model hotspot , and the distinctive correlations there can be understood by supposing that @xmath122 ; that is , the synchrotron radiation field is dominated by local emissions in the hotspot , as we might expect . using again the emissivity relations in 2.3 we can write in this case that @xmath123 and @xmath124 . taking @xmath125 for the injection model hotspot radio spectral index
we would predict @xmath126 and @xmath127 , very close to what is observed . finally , we note that on the whole ssc x - ray emission is no more tightly correlated with the radio intensity than is the ic/3k intensity , despite the fact that it is a consequence of the radio emission itself .
that comes from the fact that the ssc emissivity reflects radio emissions throughout the source , not just the local radio emission .
small regions of high radio brightness embedded in more diffuse emission tend to wash out the ssc / radio correlation , because they act like internal point sources that anisotropically illuminate their neighborhoods .
if the surrounding emission is particularly diffuse , such regions may contribute flux at considerable distances throughout the source .
that effect is particularly striking in the injection model , where the very bright radio hotspot is compact and surrounded by more diffuse radio emission .
as mentioned in the introduction , total synchrotron intensities are commonly applied in two ways to estimate magnetic field strengths in optically thin synchrotron sources .
this is not straightforward , of course , since synchrotron intensity from an optically thin source alone can not determine the strength of the source magnetic field .
some additional information must be supplied .
the more traditional and still most - common method minimizes the total energy ( or sometimes pressure ) for the relativistic particles and magnetic field , or what is close to the same thing , an equipartition is assumed between them .
a second , and increasingly common method combines information carried in the synchrotron intensity with observed inverse - compton scattered intensities presumably produced by the same electron population .
there are serious uncertainties in the application of both methods , ranging from the absence of convincing theoretical arguments for the minimum energy condition in an unrelaxed system to the likelihood of inhomogeneous magnetic field and particle distributions . in this section
we will apply standard methods for these two approaches to synthetic observations of our simulated objects in order to see how they compare and to understand better how the inferred field properties match actual conditions for objects with complex , quasi - realistic structures . for this exercise
we `` observed '' synchrotron emission at @xmath128 ghz and ic/3k emission at 1.2 kev in all three models . to explore the connections between
observationally inferred magnetic fields and actual magnetic field distributions we selected six lines of sight ( los ) for close analysis .
those los are marked in figure [ regions.f ] .
two of the los pass through lobe structures ( * los 1 * , * los 4 * ) .
three penetrate the jet ( * los 3 * , * los 5 * , * los 6 * ) and one centers on the dominant hotspot ( * los 2 * ) . *
los 4 * allows us to look at the influence of particle aging on the analyses , since the spectrum there is strongly convex in the cooling model .
similarly , * los 5 * shows a concave spectrum in the injection model , due to strong blending of dissimilar power - law electron distributions introduced at shocks .
( these spectra along * los 4 * and * los 5 * are shown in figure [ losspec.f ] . )
* los 6 * passes through the jet base , where magnetic field and particle populations are particularly simple , and known from the jet inflow conditions .
magnetic field strength can be calculated very simply from synchrotron and ic/3k intensities for a uniform medium when the electron distribution is a power - law .
we used the following expression for this estimated field , @xmath0 ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) : @xmath129 for uniform particle and field distributions , this expression is exactly equivalent to inverting the analytic calculation of synchrotron and ic/3k surface brightnesses in our synthetic observations .
here @xmath131 is the observed radio flux density within an observing beam , @xmath132 , @xmath133 is a magnetic field volume filling factor , @xmath134 is the angle between the magnetic field and the line of sight , @xmath135 and @xmath136 are the semimajor and semiminor axes of the observing beam in arcseconds , @xmath113 is the path length through the source in kpc , and @xmath137 and @xmath138 are the _ fixed _ lower and upper synchrotron cutoff frequencies in the source frame , expressed in ghz .
note that while @xmath1 is expressed in terms of synchrotron flux density , it really depends on the mean @xmath139 ; that is , the mean synchrotron emissivity .
we assume below for simplicity that @xmath140 .
the magnetic field is so tangled and intermittent that assuming @xmath141 introduces errors of only a few percent into our analysis . in order to incorporate the full nonthermal electron distributions as well as possible into @xmath1 ,
the frequency limits needed in equation ( [ bme ] ) correspond in each model to the characteristic synchrotron frequencies of the lowest and highest energy electrons and the fiducial magnetic field for each model ( see table [ summary.t ] ) .
those turn out to be @xmath142 hz and @xmath143 ghz for the control model and the injection model ; for the cooling model @xmath144 hz and @xmath145 ghz . our conclusions do not depend on any of these parameter choices .
we note that since both empirical field measures depend only on intensities , they are independent of distance assumptions .
nevertheless , for consistency with our discussions in tjr01 we place the objects at a distance of 100 mpc , giving 0.28 arcseconds for the projected size of a numerical resolution element .
summaries of our analysis are given in table [ bic1.t ] and figure [ bcomb.f ] . at high resolution
, @xmath0 generally falls within a factor of two of @xmath146 in the sampled volume .
@xmath1 , on the other hand , can show considerably larger scatter , over an order of magnitude in some cases , depending on the specifics of the particle transport .
both estimates are sensitive to spectral curvature .
estimates for jet structures tend to show less scatter than for lobe structures , which typically embody a wider range of physical conditions .
results for 3.0 and 22.0 arcsecond gaussian beams are also included in table [ bic1.t ] .
the influences of bigger beams are predictable in that derived fields resemble the analogous properties of the larger regions surveyed .
figure [ bcomb.f ] compares @xmath0 and @xmath1 to @xmath146 .
it is evident that both empirical magnetic field estimates correlate roughly with @xmath146 along the selected lines of sight .
the bolometric synchrotron intensity does depend on the rms magnetic field along the line of sight , of course .
the spectral emissivity @xmath147 , is similar , since in our sources @xmath148 are common .
still it is not obvious a priori how well such a simple measure as @xmath146 should compare to inferred values , so the experiment is valuable .
two - thirds of the @xmath0 ( @xmath1 ) points are within a factor of two ( three ) of @xmath146 . on the other hand , while the @xmath0 values are approximately randomly distributed with respect to @xmath146 , there are obvious biases in @xmath1 that are dependent on the electron transport model in the simulation .
those biases correctly reflect actual deviations from the minimum energy condition .
in particular , our simulated objects are not in electron / magnetic field energy equipartition . nor , as we will discuss in [ bic : bme ] , is there physics in the simulations expected to produce this kind of equipartition .
magnetic field profiles for each of the los are shown in figure [ blos1.f ] for the control model and injection model .
the field distributions in the cooling model are identical to the other two sources except for being everywhere a factor 14.8 times larger . however , since the particle populations evolve differently in the cooling model , the inferred magnetic field behaviors are not necessarily the same as in either of the other two models .
points where @xmath149 are set to zero in this analysis ; that is , we filtered out plasma that did not originate in the jet in order to match the assumptions made in computing nonthermal emissions . in this
regard we comment that the contact discontinuity between the `` jet plasma '' and the `` ambient plasma '' ; i.e. , the boundary between @xmath150 and @xmath151 , is not simple .
it is actually quite convoluted , so that distinct fingers of ambient plasma penetrate into the cocoon .
consequently , there are `` dropouts '' in the emission and associated physical variables according to the prescription we have followed ; that is , these regions do not contribute to @xmath0 , @xmath1 , nor to @xmath152 or @xmath146 . in [ bic : bme
] we will explore how effectively @xmath0 and @xmath1 can be used to reveal information about the partitioning of energy between magnetic fields and nonthermal electrons . here
we focus on the two field estimators separately .
we summarize the important features in the next subsection . in this section
we look directly at the properties of the magnetic field along the individual selected los as revealed in figure [ blos1.f ] .
* los 1 * and * los 4 * pass through the diffuse lobe of the sources . both lines exhibit very complex magnetic fields , and considerable entrainment of ambient plasma into the lobe structure is evident from the large number of `` dropouts '' in figure [ blos1.f ] .
the strongest field regions are revealed as the relatively bright filaments in figures [ xray - m8.f]-[xray - m7.f ] and figure [ regions.f ] ( see also fig . 5 in tjr01 ) .
the nonthermal electron density also exhibits much fine structure ( not shown ) , as one would expect from the discussion in [ bic : rad - x ] . while the ic/3k - radio analysis ( [ bic : rad - x ] ) indicates the field and particle structures
are correlated , it also shows very wide scatter , especially in the lobes , where flows are chaotic .
that does impact on @xmath0 as a quantitative field measure .
except for one case ( * los 4 * in the cooling model ) values of @xmath0 along these lobe los all fall below @xmath146 .
this bias results from the presence of substantial electron populations in weak field regions that contribute little to the radio emission , but that do produce x - rays . in effect the ic/3k intensity provides an overestimate of the number of radio emitting electrons , so that under the uniform source hypothesis the field apparently needed to account for the radio emissions is weakened . on the other hand ,
a comparison with figure [ blos1.f ] shows that @xmath0 for the control model and injection model are representative of the field values being sampled along both los , so it does give a `` meaningful '' result , if not a simply defined quantitative one .
the relatively higher value of @xmath0 for the cooling model on * los 4 * is a consequence of spectral aging ; that is , the spectrum is convex .
intentionally , no correction was made for this effect , in order to expose its potential influence .
recall that the 1.2 kev ic/3k emission used to compute @xmath0 comes from 1 gev electrons , while the 2.9 ghz emission is produced by electrons with energies about an order of magnitude higher , even in the strongest field regions along this los . with a convex electron spectrum the computed @xmath0 will be artificially increased , as simple arguments can show .
suppose , for example , we measured the bolometric synchrotron intensity , @xmath153 , from electrons of energy @xmath154 and the bolometric ic/3k intensity , @xmath155 , from electrons of energy @xmath156 . if the object were homogeneous , and we fixed the observed radio band , it is simple to show that @xmath157 , where @xmath158 and @xmath159 represent the number of electrons required to produce the observed intensities .
if we assumed that @xmath158 and @xmath159 were connected by a powerlaw in @xmath160 with the powerlaw index determined by the high energy electrons responsible for the radio band , but the distribution is actually convex , we would always overestimate @xmath161 , since we would overestimate @xmath162 . this same influence has an even more striking impact on the value of @xmath1 for the cooling model on * los 4*. under the equipartition that accompanies the minimum energy assumption , the effective overestimate of the electron population also exaggerates the estimated magnetic field energy , leading to an estimated magnetic field about two orders of magnitude greater than @xmath146 and an order of magnitude greater than @xmath152 along this los . thus , it is very important when analyzing sources that exhibit `` aged '' spectra to account for the curvature when calculating the electron energy ( i. l. tregillis and l. biggs , in preparation ) .
the values of @xmath1 found in the injection model along these los , by contrast are almost an order of magnitude below the rms field values .
this correctly reflects the fact that the magnetic field energy in this model actually does greatly exceed the nonthermal electron energy in most locations that generate synchrotron emission . * los 3 * , * los 5 * and * los 6 * intersect the jet .
@xmath0 values are close to @xmath146 in each model when the highest resolution is used in the observations . for the control model and cooling model
the inferred field is almost independent of angular resolution .
the injection model jet is relatively less dominant in accounting for emissions , since the radiating electron population is small there . at high resolution
the inferred field is also close to @xmath146 .
the concave property of the synchrotron spectrum on * los 5 * in the injection model has at most a modest depressing influence on @xmath0 . as beam size increases on * los 3 * and *
los 5 * , @xmath0 in the injection model increases markedly , because emissions become influenced by the nearby hotspot , where the electron population increases dramatically . that influence is greater in the injection model since its hotspot is much more intense than in the other two models ( see fig .
[ x1vi.f ] ) .
@xmath1 values follow a similar pattern to those seen in @xmath0 for lobe los ; that is , the relation between @xmath1 and @xmath146 follows the actual relationship between nonthermal particle and magnetic field energies in the predominant emission regions .
* los 6 * passes through the jet near its origin . in figure [ blos1.f ]
the jet is contained roughly between position coordinates 245 - 295 , measured in computation zones .
the magnetic field structure is simple and approximates that introduced at the computational boundary .
indeed , for all three models @xmath0 is very close to @xmath146 as well as the projected axial base jet magnetic field in the simulation . *
los 2 * passes through a hotspot resulting from a shock produced as the jet impinges on the near boundary of its cocoon .
the magnetic field is relatively compressed and organized there , accounting for the dominant peak in the field distribution near line coordinate 200 .
virtually all of the radio emission in this direction originates in the hotspot , so that the @xmath0 and @xmath1 values for all three models lie close to @xmath146 , irrespective of angular resolution .
this is the only los we sampled that returns a @xmath1 estimate in the injection model that is close to @xmath146 .
it is , in fact , the only los we sampled that has emissions in that electron transport model predominantly from regions close to electron / magnetic field equipartition .
in the previous section we found that a synchrotron / inverse - compton analysis provides reasonable estimates of characteristic magnetic fields , in particular @xmath146 , in our simulated objects .
there are biases that can degrade the estimates when the fields are significantly more intermittent than the nonthermal electrons or when the electron energy spectra are strongly aged . however , the @xmath0 values we derived were mostly within a factor of two of the rms fields .
on the other hand the minimum energy magnetic fields were close to the rms fields ( or any other simple statistical measure ) only when the emission was dominated by relativistic plasma close to equipartition .
that is what we should hope for and expect , of course . at the same time
, we noted that the biases in @xmath1 correctly reflected the degree of deviation from equipartition of the dominant plasma .
this suggests that combining @xmath0 and @xmath1 might reveal a meaningful estimate of the ratio of nonthermal electron and magnetic field energy and , consequently , provide an improved estimate of the total nonthermal energy compared to the minimum energy .
we test that possibility in this section for our simulated objects , where we not only can perform radio and x - ray observations , but also know the internal energetics directly .
the obvious parameter for comparing magnetic and nonthermal particle energies is the ratio of their local densities ; namely , @xmath163 standard expressions for the minimum - energy magnetic field , @xmath1 , ( i.e. , assuming a uniform source and a power - law @xmath6 , and a fixed frequency band ) ( see eq . [ [ bme ] ] ) , lead to the simple relation @xmath164 where @xmath161 is the actual field strength . the exact value of @xmath165 makes no substantial difference to our conclusions , so for convenience we will still apply the commonly used value @xmath166 .
the total energy density in nonthermal particles and magnetic fields , @xmath167 , is then obviously @xmath168~ , \label{utot.e}\end{aligned}\ ] ] where @xmath169 is the combined minimum energy density .
it is convenient to define @xmath170 corresponding to @xmath171 . for @xmath166
this gives @xmath172 .
using @xmath173 we can write simply @xmath174 or @xmath175 note for the discussion below , when @xmath176 ( @xmath177 ) , that @xmath178 ( @xmath179 ) . to set the stage for our discussion
we note that the jet inflow boundary conditions discussed in [ bic : jetprop ] give @xmath180 for the control model and the cooling model .
thus , these two jets enter the grid relatively close to electron / field equipartition .
accordingly , from equation ( [ dbme3.e ] ) one finds at the jet orifice that @xmath181 .
in contrast , the injection model jet enters very far from equipartition , with @xmath182 , so @xmath183 .
how should we expect @xmath184 to change from these input values as the flows evolve ? recall that in our simulations particle energies are dominated by the thermal population ; the nonthermal electrons are `` test particles '' that respond to the underlying flow dynamics , which is based on otherwise fully consistent mhd .
the expected variation of @xmath184 under adiabatic compression in those flows is straightforward to estimate .
recall from [ bic : rad - x ] that compression of an isotropic magnetic field varies @xmath104 , so that @xmath185 .
we found in [ bic : rad - x ] that this behavior seems to account for much of the field structure in the lobe . at the same time , the nonthermal electron gas is relativistic , so @xmath186 , as well .
thus , we may expect adiabatic expansion or compression to produce relatively little change in @xmath184
. on the other hand , in sheared flows @xmath161 may have little or no correlation with @xmath187 .
in particular , an incompressible stretched flux tube has a field strength depending only on the length of the tube ( see [ bic : rad - x ] ) , which could produce very large increases in @xmath184 inside flux tubes .
variations of @xmath184 across shocks are also difficult to quantify , although qualitatively we can expect @xmath184 to decrease there , perhaps by a large factor . at most @xmath188 ( in a perpendicular shock ) , but @xmath189 should increase more strongly than this in response to diffusive shock acceleration .
as shock strength increases the jump in @xmath190 increases asymptotically , but the jump in @xmath189 does not , both due to the diffusive reacceleration of the incident particle flux and also to any fresh injection due to `` thermal leakage '' at the shock .
because of initial conditions , most regions within the control model and cooling model _ coincidentally _ lie within an order of magnitude of this equipartition condition , with nonthermal electron energy being typically somewhat greater than magnetic field energy ( see figs .
[ dlos - m8.f],[dlos - m7.f ] ) .
in contrast for much of the source volume in the injection model , magnetic field energy greatly exceeds the nonthermal electron energy density ( fig .
[ dlos - m6.f ] ) . that contrasting condition simply reflects the small population of nonthermal electrons entering with the jet in this model .
it conveniently provides an excellent opportunity to study how different energy balances between particles and fields reveal themselves through their emission properties .
we again use the selected * los 1 - 6 * to explore how energy partitioning inferred from observations compares to actual physical conditions within the sources .
table [ dtable.t ] lists the rms and mean values , @xmath191 and @xmath192 , along each line of sight , as well as the value of @xmath193 inferred from the ratio @xmath194 used in equation ( [ dbme2.e ] ) , with @xmath140 .
the direct values of @xmath184 were computed from the magnetic field strength in each numerical bin and the nonthermal electron energy density integrated between @xmath195 and @xmath196 , corresponding to the synchrotron cutoffs assumed in calculating @xmath1 .
figure [ epart.f ] shows graphically the relationship between the @xmath193 and @xmath192 data in table [ dtable.t ] . for comparison , figures [ dlos - m8.f]-[dlos - m7.f ] display distributions of @xmath184 along * los 1 - 6 * using the actual , full nonthermal electron distribution and the magnetic field properties from the simulations .
average values , @xmath197 , are indicated in those plots by the dotted lines and @xmath193 by dashed lines .
first looking at figure [ epart.f ] we see a rough correspondence between @xmath193 and @xmath192 for most of the data . on the other hand agreement
is not generally better than about an order of magnitude . from equation ( [ utot.e ] ) that would correspond to an uncertainty in the total energy of about a factor of four , if @xmath198 were known precisely .
we should not expect exact correspondence , since @xmath193 is based on two indirect measures weighted to different points along each line of sight .
we note below a couple of evident patterns that are useful to examine in order to obtain better insights to the limits of information contained in @xmath193 . the two lobe los ( * los 1 * and * 4 * ) in the cooling model give @xmath193 values two and four orders of magnitude , respectively , below @xmath192 .
this strong bias is once again due to the influence of spectral aging in the lobes of this model , which have not been corrected for in this simple experiment .
that is , if the electron spectrum is convex , one must be careful to count properly the low energy electrons , or @xmath1 will be seriously over estimated , which has an even larger impact on @xmath193 .
most of the other los in the control model and the cooling model scatter reasonably around the @xmath199 line within about a factor of 4 , and examination of figures [ dlos - m8.f ] and [ dlos - m7.f ] shows that @xmath193 is usually `` representative '' . in that regard
it is worth noting along the jet * los 3 * , * 5 * and * 6 * that @xmath193 agrees well with physical values of @xmath184 in the jet itself , which is the dominant emission source . on the other hand @xmath193 for the control model hotspot ( * los 2 * ) misses the actual energy partitioning in the hotspot by about a factor of seven . here
the field geometry is more complex , so that @xmath1 is an under estimate of @xmath161 in the hotspot , even though it exceeds @xmath146 along the los .
all of the @xmath193 values in the injection model fall significantly below @xmath192 .
the closest match comes from * los 6 * , which is dominated by the ( simple ) jet base .
otherwise , in this model there are very large fluctuations in @xmath184 that do not correspond well to the strongest emission regions .
so , in this model @xmath192 is biased upwards compared to values representative of the emitting plasma and that determine @xmath193 . the estimation of the local energy partitioning @xmath184 based on @xmath0 and @xmath1 works best in regions that are not strongly influenced by radiative aging and where the partitioning is not expected to vary strongly over short distances , such as inside jet structures
there is increasing evidence that the x - ray emission observed from some jets is doppler - boosted ic/3k emission ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) . in that case
this kind of analysis could be useful still , but it would require taking beaming and boosting effects into account , which we have not done because the flows here are nonrelativistic .
that obviously adds another level of uncertainty , since those corrections can be large .
our analysis here was directed at the simpler case , but one that still contains many real world features that must be accepted in any complex synchrotron source .
we have performed extensive synthetic radio and x - ray observational analyses of the numerically simulated radio galaxies introduced in tjr01 .
these are the first synthetic observations with sufficient detail to allow the application of standard observational techniques to numerical simulations of radio galaxies .
standard observational techniques were applied to the simulation data in order to understand better how these techniques recover physical properties of observed objects .
the simulated objects have the advantage of known physical properties and evolve with many fewer simplifying assumptions than required in analytic studies of this kind .
we emphasize that our goal was to compare observed properties with known physical properties , not to present the simulated objects as direct models for real objects .
our models were intentionally idealized to isolate important nonthermal particle transport behaviors .
we concentrated in this paper on magnetic field strength and source energy content calculations derived from synchrotron and inverse - compton intensities .
an analysis of information extracted through polarimetry will be presented in a forthcoming companion paper .
we enumerate some of the practical messages from our analysis : 1 .
the synchrotron to inverse - compton intensity ratio provides a reasonable tool for estimating magnetic field strengths in complex radio sources when the inverse - compton photons come predominantly from the cmb . in our synthetically - observed simulated radio galaxies the standard radio / x - ray analysis returned magnetic field values that fell within about a factor two of the rms field in a sampled volume , unless the electron spectrum was strongly convex ; i.e. , in the absence of strong aging . strongly aged spectra return field values that are too high , so need to be corrected for that effect .
the effectiveness of this tool seems to apply even in regions where the electron distribution and the field structure are spatially intermittent ; i.e. , where they have small filling factors , such as in the lobes of the simulated radio galaxies .
it is largely independent of the relative partitioning of energy between electrons and magnetic field .
the standard synchrotron emission minimum energy analysis returns magnetic field values consistent with the rms fields only when the actual energy partitioning between electrons and fields is close to equipartition . otherwise the inferred fields are biased in directions that correctly reflect the actual deviations from equipartition in the regions being sampled .
in addition , our analysis demonstrated that minimum energy estimates based on strongly aged spectra can seriously overestimate the actual mimimum energy magnetic field , unless the spectral curvature is properly accounted for .
the energy partitioning and the total source energy can be roughly estimated utilizing the ratio of the minimum energy magnetic field to that inferred from the relative synchrotron and inverse - compton intensities . in our analysis
the actual energy content was recovered to within a little better than an order of magnitude , once again in the absence of strong spectral aging . without correction ,
however , total energy contents were overestimated by much more than an order of magnitude in regions with emissions dominated by strongly aged electron spectra .
4 . if sources are well resolved , it may be practical to examine correlations between the radio and x - ray intensity distributions as a probe of dynamical relationships between the particles and the magnetic field . in our synthetic observation analysis of these distributions in the simulated objects we were able to recover correctly the physical correlation between magnetic field and plasma density present in the objects .
5 . in regions with very large intensity contrasts , such as near a bright hotspot ,
smoothing at low resolution naturally leads to biases of the inferred properties in the direction of those physically in the dominant emission regions . finally , a note on relativistic effects .
our surface brightness computations would have to be modified in order to obtain meaningful numbers for the case of relativistic flows .
however , our main purpose has not been to study the details of the surface brightness distributions so much as it has been to try to understand how well we recover meaningful information from the observations if we model them in simple but appropriate ways .
it is difficult and probably of only limited usefulness to attempt a direct comparison to the relativistic case without genuine calculations .
a valuable extension of this work would be to apply the same kind of analysis to appropriately - modeled emission ( i.e. , including beaming and boosting effects ) from relativistic flows .
the work by i. l. t. and t. w. j. was supported by the nsf under grants ast96 - 16964 and ast00 - 71176 and by the university of minnesota supercomputing institute .
the work by d.r . was supported in part by kosef through grant r01 - 1999 - 00023 .
we gratefully acknowledge larry rudnick for many helpful comments and discussions , the assistance of william ryan in developing data analysis tools , and the anonymous referee for numerous valuable suggestions .
the los alamos national laboratory strongly supports academic freedom and a researcher s right to publish ; therefore , the laboratory as an institution does not endorse the viewpoint of a publication or guarantee its technical correctness .
assuming an isotropic emissivity , @xmath200 , and an optically thin intervening plasma , the omnidirectional synchrotron flux at a given position , @xmath201 , is estimated as @xmath202 on a uniform grid of zone size @xmath203 , the above equation is written as @xmath204 where @xmath205 here , @xmath206 , @xmath207 , and @xmath8 are the numbers of grid zones along the @xmath208 , @xmath209 , and @xmath210-directions , respectively .
then , from the convolution theorem , the omnidirectional flux can be computed as @xmath211 , \label{app4.e}\ ] ] where @xmath212 and @xmath213 are the fourier transform of @xmath214 and @xmath215 , respectively , and @xmath216 denotes the inverse fourier transform . for @xmath212
, we use @xmath217^{1/2}}. \label{app5.e}\ ] ] we note that the discrete fourier transform can be used only with a periodic grid , while ours is not . however , as described , for instance , in @xcite , we can make @xmath214 and @xmath215 periodic by doubling our grid along each direction ( that is , by extending the range of the summation into @xmath218 , @xmath219 , and @xmath220 in equation [ app2.e ] ) .
however , since the synchrotron emissions are generally confined well inside the computational grid , aliasing is not significant .
we have established empirically that there was little difference between using the doubled grid and using the original grid with periodic extensions .
model & i d & in - flowing & shock injection & cooling & @xmath33 + & & electrons ( @xmath221 ) & parameter ( @xmath13 ) & time ( @xmath222 ) & ( @xmath223 ) + 1 ....... & control & @xmath224 & @xmath225 & @xmath226 & @xmath227 + 2 ....... & injection & @xmath228 & @xmath224 & @xmath226 & @xmath227 + 3 ....... & cooling & @xmath224 & @xmath225 & @xmath229 & @xmath230 + field & beam & 1 & 2 & 3 & 4 & 5 & 6 + @xmath152 & & 0.56 & 1.7 & 0.88 & 0.49 & 0.82 & 0.79 + @xmath146 & & 0.24 & 0.45 & 0.30 & 0.17 & 0.35 & 0.30 + @xmath0 & 0.28 & 0.15 & 1.1 & 0.39 & 0.094 & 0.43 & 0.31 + & 3.0 & 0.13 & 0.91 & 0.40 & 0.088 & 0.43 & 0.31 + & 22.0 & 0.18 & 0.37 & 0.34 & 0.19 & 0.36 & 0.29 + @xmath1 & 0.28 & 0.37 & 1.1 & 0.64 & 0.20 & 0.68 & 0.58 + & 3.0 & 0.30 & 0.91 & 0.65 & 0.19 & 0.70 & 0.58 + & 22.0 & 0.31 & 0.45 & 0.45 & 0.36 & 0.46 & 0.44 + & & + field & beam & 1 & 2 & 3 & 4 & 5 & 6 + @xmath152 & & 0.56 & 1.7 & 0.88 & 0.49 & 0.82 & 0.79 + @xmath146 & & 0.24 & 0.45 & 0.30 & 0.17 & 0.35 & 0.30 + @xmath0 & 0.28 & 0.068 & 0.65 & 0.23 & 0.077 & 0.15 & 0.30 + & 3.0 & 0.052 & 0.69 & 0.89 & 0.087 & 1.5 & 0.29 + & 22.0 & 0.95 & 0.89 & 1.2 & 0.15 & 1.1 & 0.16 + @xmath1 & 0.28 & 0.044 & 0.42 & 0.077 & 0.020 & 0.050 & 0.044 + & 3.0 & 0.049 & 0.39 & 0.11 & 0.023 & 0.39 & 0.060 + & 22.0 & 0.13 & 0.23 & 0.23 & 0.037 & 0.24 & 0.13 + field & beam & 1 & 2 & 3 & 4 & 5 & 6 + @xmath152 & & 8.3 & 25.3 & 13.1 & 7.3 & 12.1 & 11.7 + @xmath146 & & 3.5 & 6.7 & 4.5 & 2.6 & 5.3 & 4.5 + @xmath0 & 0.28 & 2.3 & 22.6 & 6.5 & 7.9 & 8.1 & 4.6 + & 3.0 & 2.8 & 17.3 & 6.5 & 5.5 & 8.0 & 4.8 + & 22.0 & 4.0 & 7.0 & 5.6 & 8.3 & 6.2 & 2.4 + @xmath1 & 0.28 & 15.3 & 24.2 & 14.2 & 242 & 16.6 & 10.2 + & 3.0 & 12.2 & 17.3 & 11.0 & 29.8 & 12.3 & 8.8 + & 22.0 & 5.2 & 8.5 & 7.7 & 5.5 & 8.1 & 4.4 + cccccc model & i d & los & @xmath191 & @xmath192 & @xmath231 + 1 & control & 1 & 0.351 & 0.280 & 6.36 @xmath232 + & & 2 & 0.213 & 0.173 & 1.50 + & & 3 & 0.311 & 0.239 & 0.265 + & & 4 & 0.756 & 0.432 & 0.107 + & & 5 & 0.268 & 0.247 & 0.302 + & & 6 & 0.797 & 0.364 & 0.167 + 2 & injection & 1 & 148 & 96.2 & 6.88 + & & 2 & 285 & 129 & 6.92 + & & 3 & 1110 & 716 & 69.1 + & & 4 & 821 & 479 & 168 + & & 5 & 1070 & 707 & 70.1 + & & 6 & 4340 & 1970 & 1240 + 3 & cooling & 1 & 0.278 & 0.217 & 1.98 @xmath233 + & & 2 & 0.783 & 0.539 & 1.18 + & & 3 & 0.584 & 0.373 & 9.73 @xmath232 + & & 4 & 0.192 & 0.141 & 9.43 @xmath234 + & & 5 & 0.276 & 0.226 & 0.122 + & & 6 & 1.170 & 0.413 & 9.24 @xmath232 + | we present an extensive synthetic observational analysis of numerically - simulated radio galaxies designed to explore the effectiveness of conventional observational analyses at recovering physical source properties .
these are the first numerical simulations with sufficient physical detail to allow such a study .
the present paper focuses on extraction of magnetic field properties from nonthermal intensity information .
synchrotron and inverse - compton intensities were effective in providing meaningful information about distributions and strengths of magnetic fields , although considerable care was called for in quantitative usage of the information .
correlations between radio and x - ray surface brightness correctly revealed useful dynamical relationships between particles and fields , for example .
magnetic field strength estimates derived from the ratio of x - ray to radio intensity were mostly within about a factor of two of the rms field strength along a given line of sight .
when emissions along a given line of sight were dominated by regions close to the minimum energy / equipartition condition , the field strengths derived from the standard power - law - spectrum minimum energy calculation were also reasonably close to actual field strengths , except when spectral aging was evident .
otherwise , biases in the minimum - energy magnetic field estimation mirrored actual differences from equipartition .
the ratio of the inverse - compton - estimated magnetic field to the minimum - energy magnetic field provided a rough measure of the actual total energy in particles and fields in most instances , although this measure was accurate within only about an order of magnitude .
this may provide a practical limit to the accuracy with which one may be able to establish the internal energy density or pressure of optically thin synchrotron sources . |
consider @xmath2 a subset of a group @xmath3 .
the elements of @xmath2 constitute the vertices of the commuting graph @xmath4 , in which @xmath5 are joined by an edge whenever @xmath6 .
if @xmath2 is a set of involutions then we call @xmath4 a commuting involution graph .
commuting graphs have been studied by many authors in various contexts .
for example fischer @xcite studied commuting involution graphs for the case when @xmath2 is a conjugacy class of involutions , in his work on 3-transposition groups .
segev and seitz looked in @xcite at commuting graphs for finite simple groups @xmath3 where @xmath2 consists of the non - identity elements of @xmath3 .
more recently giudici and pope @xcite gave some results on bounding the diameters of commuting graphs of finite groups .
commuting graphs for elements of order 3 have been considered in @xcite ; there have also been many papers dealing with the case where @xmath2 consists of all non - identity elements of a given group such as for example @xcite .
+ in @xcite , bates et al looked into the commuting involution graph @xmath4 where @xmath2 is a conjugacy class of involutions in @xmath3 and @xmath3 is @xmath7 .
the remaining finite coxeter groups were analysed in @xcite .
commuting involution graphs in affine coxeter groups of type @xmath8 have been considered in @xcite . in this article , we investigate the commuting involution graphs @xmath9 for type @xmath10 . + for the rest of this article , let @xmath11 be an affine weyl group of type @xmath12 , for some @xmath13 , writing @xmath3 when @xmath14 is not specified , and let @xmath2 be a conjugacy class of involutions of @xmath3 .
we write @xmath15 for the diameter of @xmath4 when @xmath4 is a connected graph , in other words the maximum distance @xmath16 between any @xmath5 in the graph .
since conjugation by any group element induces a graph automorphism , we can determine the diameter by fixing any @xmath17 , and then @xmath18 .
our main result is the following .
[ summary ] let @xmath2 be a conjugacy class of involutions in the group @xmath11 of type @xmath12 . if the commuting graph @xmath4 is connected , then its diameter is at most @xmath1 .
the proof of theorem [ summary ] is broken into several cases depending on the type of the conjugacy class . to describe this further , and in order to state our results about connectedness , we need to describe a parameterisation of the involution conjugacy class .
we will explain in section 2 that involutions in @xmath3 can be written as ` labelled permutations ' ( notation [ notation ] ) .
these are permutations expressed as products of disjoint cycles in which every cycle has a sign and an integer written above it .
for example in @xmath19 one of the involutions is @xmath20 .
a cycle is positive if it has a plus sign , and negative if it has a minus sign .
the _ labelled cycle type _ of an involution @xmath21 will be a quadruple @xmath22 where @xmath23 is the number of transpositions , @xmath24 is the number of negative @xmath25-cycles with an even number above them , @xmath26 is the number of negative 1-cycles with an odd number above them , and @xmath27 is the number of positive 1-cycles . for the example in @xmath19 ,
its labelled cycle type is @xmath28 .
+ we will show ( theorem [ conjclass ] ) that involution conjugacy classes in @xmath3 are parametrised by labelled cycle type .
we may now give the conditions under which the graph @xmath4 is connected .
[ main ] let @xmath2 be a conjugacy class of involutions in @xmath11 ( where @xmath29 ) with labelled cycle type @xmath22 .
then @xmath4 is disconnected in each of the following cases .
@xmath30 and @xmath31 ; @xmath32 , @xmath31 and either @xmath33 or @xmath34 ; @xmath32 and @xmath35 ; @xmath36 and @xmath37 ; @xmath38 , @xmath37 and @xmath39 . in all other cases ,
@xmath4 is connected .
we note that the bound on diameter given in theorem [ summary ] is best possible .
for example we have verified that when @xmath40 , @xmath37 and @xmath41 , the commuting involution graph has diameter 10 . in section [ sec2 ]
we will establish notation and describe the conjugacy classes of involutions in @xmath3 .
section 3 is dedicated to proving the main theorems . in section 4
we give examples of selected commuting involution graphs .
for the remainder of this paper @xmath42 is a fixed involution with @xmath2 its conjugacy class in the affine irreducible coxeter group @xmath3 .
our first job is to establish what involutions in @xmath3 look like .
+ let @xmath43 be a finite weyl group with root system @xmath44 and @xmath45 the set of coroots .
the affine weyl group @xmath46 is the semidirect product of @xmath43 with translation group @xmath47 of the coroot lattice @xmath48 of @xmath43 .
see , for example , ( * ? ? ?
* chapter 4 ) for more detail . for @xmath49 and @xmath50
we have @xmath51 for the rest of the paper @xmath52 will denote a coxeter group of type @xmath53 , and set @xmath54 .
we write , respectively , @xmath43 and @xmath3 whenever it is not necessary to specify @xmath14 .
we may take the roots of @xmath43 to be @xmath55 and @xmath56 , for @xmath57 , where @xmath58 is the standard orthonormal basis for @xmath59 .
the coroots are then @xmath60 and @xmath56 . therefore in this case
@xmath61 we may view the elements of @xmath43 as signed permutations ; they act on @xmath59 by permuting the subscripts of basis vectors and changing their signs . to obtain a signed permutation we write a permutation in sym(@xmath14 ) ( including 1-cycles ) , add a plus sign or a minus sign above each @xmath62 , and say @xmath62 is positive or negative accordingly .
we adopt the convention of reading the sign first ; that is , if @xmath63 , then @xmath64 , @xmath65 and @xmath66 .
+ expressing @xmath21 as a product of disjoint cycles , we say that a cycle @xmath67 of @xmath21 is _ positive _ if there is an even number of minus signs above its elements , and _ negative _ if the number of minus signs is odd .
for example , @xmath68 is a negative cycle , whereas @xmath69 is positive .
it is straightforward to check that an involution of @xmath43 only has 1-cycles ( positive or negative ) and positive 2-cycles . by the definition of group multiplication in @xmath11
, we see that the element @xmath70 of @xmath3 is an involution precisely when @xmath71 .
this allows us to characterise the involutions in @xmath11 .
[ invn ] a non - identity element @xmath70 of @xmath11 is an involution if and only if @xmath21 , when expressed as a product of disjoint signed cycles , has the form @xmath72 for some @xmath73 and @xmath27 ; and , writing @xmath74 , we have @xmath75 when @xmath21 contains @xmath76 , @xmath77 when @xmath21 contains @xmath76 and @xmath78 for @xmath79 . if @xmath80 is an involution , then @xmath81 .
thus @xmath21 is an involution of @xmath43 , and so its cycles are all either 1-cycles or positive 2-cycles . thus @xmath21 has the form given in the statement of the lemma .
write @xmath82 .
we must have @xmath83 for all @xmath84 .
for @xmath85 , we have @xmath86 .
thus @xmath75 . for @xmath87
, we have @xmath88 and @xmath89 .
so @xmath77 .
for @xmath90 we have @xmath91 , so there is no restriction on @xmath92 .
for @xmath93 we have @xmath94 , forcing @xmath78 .
these are the necessary conditions on @xmath21 and @xmath95 ; it is clear that they are also sufficient .
[ notation ] from now on we will employ a shorthand for writing involutions @xmath96 of @xmath3 : if @xmath74 , then above each signed number @xmath62 in the expression of @xmath21 as a product of disjoint signed cycles , we will write @xmath97 . however for transpositions @xmath98 of @xmath21 , where the number above @xmath42 determines the number above @xmath99
as described in lemma [ invn ] , we write @xmath100 for @xmath101 and @xmath102 for @xmath103
. we will call this the _ labelled cycle form _ of @xmath42 . where it is helpful , we adopt the convention that cycles @xmath104 are omitted , as these fix both @xmath42 and @xmath105 .
let @xmath42 be an involution in @xmath11 .
the _ labelled cycle type _ of @xmath42 is the tuple @xmath106 , where @xmath23 is the number of transpositions , @xmath24 is the number of negative 1-cycles with an even number above them , @xmath26 is the number of negative 1-cycles with an odd number above them , and @xmath27 is the number of positive 1-cycles ( fixed points ) , in the labelled cycle form of @xmath42 .
for example , the labelled cycle type of @xmath107 is @xmath108 .
+ having characterised the involutions , we must now determine the conjugacy classes .
a well - known result , due to richardson , gives a description of involution conjugacy classes in coxeter groups .
[ equiv ] let @xmath43 be an arbitrary coxeter group , with @xmath109 the set of fundamental reflections .
we say that two subsets @xmath110 and @xmath111 of @xmath109 are @xmath43-equivalent if there exists @xmath112 such that @xmath113 . in the next result
, we use the notation @xmath114 for the longest element of a finite standard parabolic subgroup @xmath115 .
[ richardson ] let @xmath43 be an arbitrary coxeter group , with @xmath109 the set of fundamental reflections .
let @xmath116 be an involution .
then there exists @xmath117 such that @xmath118 is central in @xmath115 , and @xmath119 is conjugate to @xmath118 .
in addition , for @xmath120 , @xmath118 is conjugate to @xmath121 if and only if @xmath110 and @xmath111 are @xmath43-equivalent .
the coxeter graphs of @xmath122 and @xmath123 , @xmath124 are as follows .
+ 1.00 mm ( 58.00,20)(-10,0 ) ( 10,10.80)(1,0)12 ( 10,9.00)(1,0)12 ( 10,10.00 ) ( 22,10.00 ) ( 34,10.00 ) ( 10,14)(0,0)[cc]@xmath125 ( 22,14)(0,0)[cc]@xmath126 ( 32,14)(0,0)[cc]@xmath127 ( -10,10)(0,0)[lc]@xmath128 ( 22,10.80)(1,0)12 ( 22,9.00)(1,0)12 1.00 mm ( 58.00,14)(-30,0 ) ( 10,10.80)(1,0)12 ( 10,9.00)(1,0)12 ( 22,10.00)(1,0)8 ( 50,10.00)(1,0)8 ( 58,10.80)(1,0)12 ( 58,9.02)(1,0)12 ( 58.00,10.00 ) ( 10,10.00 ) ( 22,10.00 ) ( 70,10.00 ) ( 10,14)(0,0)[cc]@xmath125 ( 22,14)(0,0)[cc]@xmath126 ( 58,14)(0,0)[cc]@xmath129 ( 70,14)(0,0)[cc]@xmath130 ( -20,10)(0,0)[lc ] @xmath131 ( 35.00,10.00)(1,0)0.4 ( 45.00,10.00)(1,0)0.4 ( 40.00,10.00)(1,0)0.4 we may set @xmath132 , @xmath133 for @xmath134 , and @xmath135 . + it is well known that in the finite coxeter group @xmath43 of type @xmath53 , elements are conjugate if and only if they have the same signed cycle type . in particular two involutions are conjugate when they have the same number of transpositions , the same number of negative 1-cycles and the same number of positive 1-cycles . + in @xmath3 ( which is of type @xmath12 ) , the element @xmath136 is conjugate to @xmath137 via some @xmath138 if and only if : @xmath139 [ conjclass ] involutions in @xmath3 are conjugate if and only if they have the same labelled cycle type .
in particular , every involution is conjugate to exactly one element @xmath140 of the form @xmath141 let @xmath142 be an involution of @xmath3 . by theorem [ richardson ] , @xmath142 is conjugate to @xmath118 for some finite standard parabolic subgroup @xmath115 of @xmath3 in which @xmath118 is central .
therefore the connected components of the coxeter graph for @xmath115 are of types @xmath143 or @xmath144 for some @xmath62 ( including , by a slight abuse of notation , @xmath145 , where we have connected components with just the vertex @xmath146 or @xmath130 ) .
thus @xmath147 for some @xmath148 with @xmath149 , where @xmath111 is a subset of @xmath150 no two elements of which are adjacent vertices in the coxeter graph .
by conjugation in @xmath151 ( which after all is isomorphic to the symmetric group @xmath152 ) , we can assume that for some @xmath148 and @xmath23 with @xmath153 we have @xmath154 this gives that @xmath142 is conjugate to @xmath118 , where @xmath155 let @xmath156 , where @size10@mathfonts @xmath157 from equation we see that @xmath158 therefore , by setting @xmath159 we see that each involution in @xmath3 is conjugate to at least one element of the required form .
+ now consider an involution @xmath142 in @xmath3 with labelled cycle type @xmath22 , and suppose @xmath160 for some simple reflection @xmath161 .
write @xmath162 and @xmath163 . by equation
, @xmath164 is conjugate to @xmath21 in the underlying weyl group @xmath43 .
hence @xmath164 and @xmath21 have the same number of transpositions , negative 1-cycles and positive 1-cycles as each other .
in other words , the labelled cycle type of @xmath165 is @xmath166 for some @xmath167 satisfying @xmath168 .
now , from equation again , @xmath169 is a labelled 1-cycle of @xmath142 if and only if @xmath170 is a labelled 1-cycle of @xmath165 , where @xmath171 and @xmath172 .
in particular this means @xmath173 is a signed cycle of @xmath164 , so that @xmath174 .
now @xmath175 .
so @xmath176 .
therefore @xmath177 . hence @xmath178 and so @xmath142 and @xmath165
have the same labelled cycle type .
in particular @xmath142 is conjugate to at most one element @xmath42 of the form stated in the theorem .
conversely , any two involutions of the same labelled cycle type @xmath22 are both conjugate to @xmath179 , and hence to each other .
thus conjugacy is parameterised by labelled cycle type , and the set of elements @xmath180 contains exactly one representative of each conjugacy class of involutions in @xmath11 .
we next prove three preliminary lemmas which will be used repeatedly in the proofs in section [ sec3 ] .
[ 1cycle ] let @xmath181 and @xmath182 .
then @xmath183 commutes with @xmath184 for all @xmath185 , whereas @xmath183 commutes with @xmath186 if and only if @xmath187 . for convenience , we assume without loss of generality that @xmath188 .
now @xmath189 , so @xmath190 commutes with @xmath191 . in the case of two negative 1-cycles we have @xmath192 and @xmath193 , so @xmath190 commutes with @xmath194 if and only if @xmath195 .
[ 2cycle ] let @xmath196 be distinct elements of @xmath197 and let @xmath185 , @xmath198 and @xmath199 be integers .
@xmath200 and @xmath201 commute if and only if @xmath187 , and @xmath202 and @xmath203 commute if and only if @xmath187 . but @xmath200 and @xmath203 commute for all @xmath185 and @xmath198
. @xmath204 and @xmath205 commute for all @xmath185 , but there is no value of @xmath198 or @xmath185 for which @xmath204 and @xmath206 or @xmath207 commute .
@xmath208 and @xmath209 commute if and only if @xmath210 , whereas @xmath202 and @xmath209 commute if and only if @xmath211 .
we lose no generality by assuming , for ease of notation , that @xmath188 , @xmath212 and @xmath213 . recall that @xmath214 and @xmath215 are shorthand for @xmath216 and @xmath217 respectively .
involutions commute precisely when their product is an involution ( or the identity ) , which we can check using lemma [ invn ] .
we calculate @xmath218 this means @xmath214 and @xmath219 commute if and only if @xmath187 .
we also have @xmath220 so again @xmath215 and @xmath221 commute if and only if @xmath187 .
finally a similar calculation shows that @xmath222 , which is an involution for all values of @xmath185 and @xmath198 , so @xmath214 and @xmath221 always commute .
certainly @xmath223 and @xmath224 commute for all @xmath185 ; after all , @xmath224 is just the identity element when @xmath213 .
moreover , already in the underlying group @xmath43 of type @xmath225 we observe that @xmath226 does not commute with @xmath227 or @xmath228 , so there is no value of @xmath198 or @xmath185 for which @xmath223 and @xmath229 or @xmath230 commute .
we have @xmath231 ; this is an involution if and only if @xmath232 .
thus @xmath233 and @xmath234 commute if and only if @xmath210 .
similarly @xmath235 ; this is an involution if and only if @xmath236 .
therefore @xmath237 and @xmath238 commute if and only if @xmath211 . [ doubletrans ] let @xmath239 , @xmath240 , @xmath241 , @xmath242 , @xmath243 and @xmath244 , for distinct @xmath245 in @xmath197 and integers @xmath246 . then @xmath247 commutes with @xmath248 if and only if @xmath249
; @xmath247 does not commute with @xmath250 ; @xmath247 commutes with @xmath251 if and only if @xmath252 ; @xmath253 commutes with @xmath250 if and only if @xmath254 ; @xmath253 does not commute with @xmath251 ; @xmath255 commutes with @xmath251 if and only if @xmath256 .
we lose no generality by assuming , for ease of notation , that @xmath36 and @xmath257 @xmath258 and @xmath259 .
hence @xmath260 now @xmath247 and @xmath248 commute if and only if @xmath261 is an involution .
this occurs if and only if @xmath262 and @xmath263 .
rearranging gives @xmath254 , as required for part ( i ) .
since @xmath264 and @xmath265 do not commute , it is impossible for @xmath247 to commute with @xmath250 .
we have @xmath258 and @xmath266 , and @xmath267 now @xmath247 and @xmath251 commute if and only if @xmath268 is an involution .
from the above calculation this occurs if and only if @xmath269 and @xmath270 .
that is , if and only if @xmath271 .
we calculate @xmath272 .
thus @xmath253 commutes with @xmath250 if and only if @xmath273 .
since @xmath274 and @xmath275 do not commute , @xmath253 can not commute with @xmath251 .
this part follows from the fact that @xmath276 .
we end this section by stating the relevant results from @xcite about commuting involution graphs in weyl groups of type @xmath277 . [
finite ] suppose that @xmath43 is of type @xmath277 , and let @xmath278 set @xmath279 and @xmath280 .
then the following hold . if @xmath30 , then @xmath4 is a complete graph . if @xmath281 , then @xmath282 . if @xmath283 and @xmath284 , then @xmath4 is disconnected . if @xmath285 and @xmath286 , then @xmath287 .
if @xmath288 , @xmath37 and @xmath289 then @xmath290 . if @xmath288 , @xmath37 and @xmath291 then @xmath292 . finally if @xmath36 , @xmath37 and @xmath289 then @xmath4 is disconnected
we begin by looking at connectedness . for an element @xmath293 in a conjugacy class @xmath2 of @xmath3 , define @xmath294 .
then let @xmath295 be the conjugacy class of @xmath296 in @xmath43 . clearly if @xmath297 , then @xmath298 .
[ easy ] suppose @xmath297 . if @xmath299 , then @xmath300 . if @xmath301 is disconnected , then @xmath302 is disconnected .
the result follows immediately from the observation that is @xmath119 commutes with @xmath303 in @xmath3 , then @xmath296 commutes with @xmath304 in @xmath43
. we can now prove theorem [ main ] ( which gives necessary and sufficient conditions for @xmath4 to be disconnected ) in one direction .
the proof in the other direction will arise from bounding the diameters of graphs not shown in theorem [ mainhalf ] to be disconnected .
[ mainhalf ] let @xmath2 be a conjugacy class of involutions in @xmath11 ( where @xmath29 ) with labelled cycle type @xmath22
. then @xmath4 is disconnected in each of the following cases .
@xmath30 and @xmath31 ; @xmath32 , @xmath31 and either @xmath33 or @xmath34 ; @xmath32 and @xmath35 ; @xmath36 and @xmath37 ; @xmath38 , @xmath37 and @xmath39 .
let @xmath2 be a conjugacy class of involutions in @xmath11 , with labelled cycle type @xmath305 , and @xmath306 as defined in theorem [ conjclass ] .
we deal with each case in turn .
suppose @xmath31 and @xmath30 .
then any @xmath142 in @xmath2 is of the form @xmath307 for some @xmath308 ( where @xmath24 of the @xmath308 are even and @xmath26 are odd ) . by lemma [ 1cycle ]
, @xmath142 does not commute with any other element of @xmath2 .
so in fact @xmath4 is completely disconnected in this case .
suppose @xmath31 and @xmath309 ( the case @xmath33 is similar )
. then @xmath310 suppose @xmath311 such that @xmath42 commutes with @xmath99 .
consider the cycle of @xmath99 that contains @xmath14 .
this must be a negative 1-cycle or a 2-cycle , because @xmath99 has the same labelled cycle type as @xmath42 .
if it is a 2-cycle , then @xmath99 can not commute with @xmath42 , by lemma [ 2cycle](iii ) .
therefore it is a negative 1-cycle @xmath312 where @xmath185 is odd , and then by lemma [ 1cycle ] , @xmath99 contains @xmath313 .
the same argument shows that any element @xmath314 of @xmath2 which commutes with @xmath99 must also contain @xmath313 , and inductively all elements in the connected component of @xmath4 containing @xmath42 must contain @xmath313 .
therefore @xmath4 is disconnected .
suppose @xmath284 and @xmath35 .
if @xmath31 then @xmath4 is disconnected by ( ii ) .
so we can assume @xmath315 .
if either of @xmath24 or @xmath26 is zero , then theorem [ finite](iii ) and lemma [ easy ] imply that @xmath4 is disconnected .
it remains to consider the case @xmath316 . here
, @xmath317 and @xmath318 suppose @xmath311 such that @xmath42 commutes with @xmath99 , and suppose @xmath319 is a 1-cycle of @xmath99 .
if @xmath320 then as @xmath42 contains a transposition @xmath321 for some @xmath322 , lemma [ 2cycle ] ( ii ) and ( iii ) show that the only way @xmath42 and @xmath99 could commute is if @xmath99 contained @xmath323 or @xmath324 where @xmath325 , contradicting our assumptions about the labelled cycle type of @xmath99 .
therefore the elements appearing in 1-cycles of @xmath99 are @xmath326 , @xmath327 and @xmath14 .
inductively this holds for all elements in the connected component of @xmath4 containing @xmath42 . therefore @xmath4 is disconnected .
suppose @xmath36 and @xmath37 .
if @xmath35 then @xmath4 is disconnected by ( iii ) .
if this does nt happen , then one of @xmath24 , @xmath26 or @xmath27 is 2 . by theorem [ finite](v ) and
lemma [ easy ] , @xmath4 is again disconnected .
suppose @xmath38 , @xmath37 and @xmath39 . for any @xmath142 in @xmath2
, we have @xmath328 , where @xmath329 , @xmath198 and @xmath330 are even , and @xmath331 are odd .
associate a set @xmath332 to @xmath142 .
given that , by lemma [ 2cycle ] , a transposition can only commute with a pair of negative 1-cycles if either both cycles are odd or both are even , and also with reference to lemma [ 1cycle ] , we see that if @xmath165 in @xmath2 commutes with @xmath142 , then @xmath333 . therefore , for example , @xmath334 and @xmath335 are not connected in @xmath4 .
our first result bounding diameters is when @xmath30 .
[ disc ] if @xmath30 and @xmath336 , then @xmath337 in all other cases @xmath338 . in this case
we have @xmath339 let @xmath340 .
then for appropriate @xmath341 and @xmath342 we have that @xmath343 .
since @xmath344 , exactly @xmath24 of the @xmath308 must be even .
if @xmath345 , that is , @xmath346 , then @xmath119 commutes with an element @xmath303 in @xmath2 which contains the cycles @xmath347 .
now @xmath303 certainly commutes with @xmath42 , and so @xmath348 .
+ now we suppose that @xmath349 , and that one of @xmath24 and @xmath26 is zero . without loss of generality
, we may choose @xmath350 .
then , writing @xmath351 , we have @xmath352 . since conjugation by elements of centralizer of @xmath42 preserves distance in @xmath4 , without loss of generality we may take @xmath119 to be of the following form for some integer @xmath84 with @xmath353 and even integers @xmath308
: @xmath354 now consider the following sequence , where @xmath355 and @xmath356 , @xmath357 , @xmath358 are arbitrary even integers .
@xmath359 it is clear that @xmath360 commutes with @xmath361 for @xmath362 .
moreover @xmath119 commutes with @xmath363 , and @xmath364 commutes with @xmath42 . if @xmath27 divides @xmath365 , then @xmath366 , which implies that @xmath367 .
if @xmath27 does not divide @xmath365 , then @xmath368 .
the diameter of the graph in each case does equal this bound because at each stage of a path from @xmath119 to @xmath42 we can add at most @xmath27 to the number of correct negative 1-cycles , but this requires the element being considered to share no fixed points with @xmath42 .
hence , for example the element @xmath369 must be distance at least @xmath370 from @xmath42 . +
the remaining case is when @xmath24 and @xmath26 are both nonzero , and @xmath349 .
+ for any @xmath371 define @xmath372 to be the number of ` correct ' negative 1-cycles in @xmath142 .
that is , cycles @xmath373 where @xmath374 or @xmath375 where @xmath376 .
thus , for example , @xmath377 .
we say that other negative 1-cycles are ` incorrect ' .
let @xmath344 .
then @xmath119 commutes with an element @xmath378 whose positive 1-cycles are @xmath379 .
we will describe a sequence @xmath380 where at each stage @xmath381 is an element of @xmath2 such that @xmath382 and the positive 1-cycles of @xmath381 are @xmath383 for some @xmath384 where @xmath385 and @xmath386 , with @xmath387 .
moreover , for @xmath388 , @xmath381 will commute with @xmath389 . + observe that the positive 1-cycles of @xmath378 have the required form , and @xmath390 .
assume that we have @xmath391 and let the positive 1-cycles of @xmath381 be @xmath383 . to form @xmath392 we look for incorrect cycles @xmath393 of @xmath381 , where each @xmath394 is
even , each @xmath395 is odd , @xmath396 and @xmath397 .
if such cycles can be found , then @xmath381 commutes with @xmath392 where @xmath392 is given by replacing the cycles @xmath398 of @xmath381 with @xmath399 , and leaving all other cycles unchanged .
it is clear that @xmath392 commutes with @xmath381 ; moreover @xmath400 , has the appropriate positive 1-cycles and @xmath401
. this sequence can continue until we have some @xmath381 with cycles @xmath383 but ( without loss of generality ) fewer than @xmath84 incorrect cycles @xmath402 where @xmath185 is even and @xmath403 .
suppose there are exactly @xmath404 such cycles ( with @xmath405 ) .
so @xmath381 has the cycles @xmath406 . since @xmath407 for each @xmath408 there must be @xmath409 where @xmath410 for each @xmath411 , and even numbers @xmath394
such that @xmath381 has the cycles @xmath412 .
similarly @xmath381 has cycles @xmath413 for odd @xmath198 , some @xmath414 , @xmath397 and @xmath415 .
we now define @xmath165 to be @xmath416 with @xmath417 replaced with @xmath418 now @xmath381 commutes with @xmath165 , and @xmath419 .
we observe that @xmath420 .
notice that every @xmath421 with @xmath376 is either a fixed point of @xmath165 or appears in a cycle @xmath422 with @xmath198 odd .
every incorrect even negative 1-cycle of @xmath165 features a fixed point of @xmath42 , so for some @xmath423 , and conjugating @xmath165 by a suitable element of the centralizer of @xmath42 if necessary , we can assume that the even negative 1-cycles of @xmath165 are @xmath424 now @xmath165 has at least @xmath425 correct negative 1-cycles .
if we ignore the even negative 1-cycles and the correct odd negative 1-cycles , then @xmath426 cycles remain ( including @xmath27 fixed points ) .
we can use the result for @xmath427 on this remaining part of @xmath165 to see that @xmath165 is distance at most @xmath428 from the element @xmath429 of @xmath2 given by @xmath430 now @xmath429 commutes with @xmath42 , and so @xmath431 . if @xmath432 and @xmath433 , then in fact @xmath434 so @xmath435 . if @xmath436 then again @xmath435 .
if @xmath437 then we have @xmath438 . to give a lower bound on the diameter consider @xmath439 to create @xmath27 additional correct negative 1-cycles at each stage of a path from @xmath119 to @xmath42 one requires each fixed point to be a point not fixed by @xmath42 ; moreover in this case completion of the process for , say , the even negative 1-cycles requires the recreation of at least one fixed point between @xmath440 and @xmath14 and hence fewer than @xmath27 correct negative 1-cycles being created at the next stage .
thus when @xmath315 we have @xmath441 , and when @xmath437 we have @xmath442 .
this completes the proof . from now on , assume that @xmath284
. then we can take @xmath443 where @xmath444 .
[ m>0k=0 ] suppose @xmath445 .
if @xmath446 , then @xmath447 . if @xmath448 and @xmath449 then @xmath450 .
let @xmath371 .
we can write @xmath451 and @xmath452 , where @xmath453 are conjugate elements of the underlying weyl group @xmath43 . define an element @xmath454 , by @xmath455 , where @xmath456 is the unique central involution of @xmath43 .
the result is that every minus sign in @xmath303 corresponds to a plus sign in @xmath457 , and every plus sign corresponds to a minus sign . since @xmath281 , @xmath303 and @xmath457
are conjugate in @xmath43 .
so , by theorem [ finite](i ) , @xmath458 when @xmath459 and @xmath460 otherwise .
now @xmath461 and @xmath165 commutes with @xmath142 by lemmas [ 1cycle ] and [ 2cycle ] .
the result follows immediately .
[ mnonzerol=0 ] suppose @xmath462 , @xmath31 , one of @xmath24 and @xmath26 is zero , and either @xmath463 or @xmath464 ( or both ) .
then @xmath4 is connected with diameter at most @xmath14 . if @xmath371 and @xmath465 then @xmath37 and the transposition of @xmath142 is @xmath466 for some @xmath185 .
let @xmath371 .
without loss of generality @xmath350 .
we start with the case @xmath37 .
we will show that if the transposition of @xmath142 is @xmath466 for some @xmath185 , then @xmath467 .
otherwise @xmath468 .
the graph for @xmath288 ( and hence @xmath37 ) is figure [ l0 ] .
we can see from this graph that the inductive hypothesis holds for @xmath288 , so assume @xmath469 .
suppose the transposition of @xmath142 contains some @xmath470 with @xmath471 .
then by lemma [ 2cycle](iii ) @xmath142 commutes with some @xmath472 such that @xmath165 has the 1-cycle @xmath373 and such that the transposition of @xmath165 is not @xmath466 . ignoring the cycle @xmath373 we can work within @xmath473 , to see that inductively @xmath474 .
hence @xmath468 .
if the transposition of @xmath142 is @xmath466 then @xmath142 certainly commutes with an element of @xmath2 which does not have this transposition .
so @xmath467 as required .
+ now we assume @xmath475 and proceed by induction on @xmath24 to show that @xmath476 .
suppose @xmath477 .
then @xmath142 is distance at most 2 from an element @xmath165 of @xmath2 which has the transposition @xmath478 . to see this , note that if both @xmath327 and @xmath14 appear in transpositions of @xmath142 , or if both appear in 1-cycles of @xmath142 , then lemma [ 2cycle ] or lemma [ doubletrans ] , as appropriate , implies that @xmath142 commutes with some @xmath479 in @xmath2 which contains a transposition of the form @xmath480 for some @xmath185 . then @xmath479 commutes with @xmath165 .
if on the other hand , @xmath142 contains ( for example ) the 1-cycle @xmath481 and @xmath327 appears in a transposition @xmath482 for some @xmath483 less than @xmath327 , then @xmath142 commutes with @xmath484 in @xmath2 containing the transpositions @xmath485 and @xmath486 for some @xmath185 and @xmath487 .
lemma [ doubletrans ] now implies that @xmath142 commutes with an appropriate @xmath165 , in particular one containing the transpositions @xmath488 and @xmath489 .
now @xmath165 in turn commutes with some @xmath429 in @xmath2 with the 1-cycles @xmath490 and @xmath491 .
if we ignore these cycles and work in @xmath492 , then table [ tableg4 ] implies that when @xmath38 @xmath493 , and when @xmath494 proposition [ m>0k=0 ] tells us that @xmath495 . therefore @xmath476 .
finally , suppose @xmath496 and @xmath497 .
suppose there is some transposition of @xmath142 containing an element @xmath470 with @xmath498 . then by lemma [ 2cycle](iii ) @xmath142 commutes with some @xmath472 such that @xmath165 has the 1-cycle @xmath373 . by induction @xmath474 .
hence @xmath468 .
the final possibility is that the elements of the transpositions of @xmath142 are @xmath499 . since @xmath463 we can use lemma [ doubletrans ] to show that @xmath142 commutes with some @xmath165 in @xmath2 containing the transposition @xmath500 .
working in @xmath501 ( using the case @xmath37 and induction on @xmath23 ) we see that @xmath474 .
hence @xmath468 , which completes the proof of proposition [ mnonzerol=0 ] .
[ kois2 ] suppose @xmath37 , @xmath31 , @xmath502 , @xmath503 and @xmath371 .
then @xmath302 is connected with diameter at most @xmath504 .
let @xmath371 .
the distance of @xmath142 from @xmath42 will largely hinge on the whereabouts of @xmath14 and @xmath327 .
but first we deal with the case where the transposition of @xmath142 is @xmath505 . here
@xmath142 commutes with an element @xmath165 of @xmath2 having cycles @xmath506 with @xmath198 and @xmath330 odd .
using proposition [ mnonzerol=0 ] on the remaining cycles of @xmath165 , we see that @xmath165 is distance at most @xmath326 from the element @xmath99 of @xmath2 whose cycles are the same as @xmath42 except that we have @xmath507 instead of @xmath508 . clearly @xmath509 .
hence @xmath510 .
this also shows that any element with odd negative 1-cycles @xmath506 is distance at most @xmath14 from @xmath42 .
assume from now on that the transposition of @xmath142 is not @xmath505 , and that its odd negative 1-cycles are not @xmath506 .
+ if the transposition of @xmath142 is @xmath511 then @xmath142 is distance 2 from some @xmath472 with cycles @xmath512 . ignoring these cycles we use proposition [ mnonzerol=0 ] in @xmath492 to see that @xmath474 if the transposition of @xmath165 is @xmath505 and @xmath513 otherwise
thus @xmath467 . if @xmath142 has cycles @xmath514 where @xmath515 , then @xmath142 is distance 3 from some @xmath165 in @xmath2 with cycles @xmath512 .
thus @xmath510 if the transposition of @xmath142 is @xmath505 and @xmath467 otherwise .
if one of @xmath327 and @xmath14 , say @xmath14 , appears in the transposition of @xmath142 and the other appears in an even negative 1-cycle @xmath516 , then @xmath142 commutes with some @xmath517 with the cycles @xmath518 .
then @xmath479 is distance 2 from some @xmath472 with the cycles @xmath512 , and such that the transposition of @xmath165 is not @xmath519 for any @xmath21 . by proposition [ mnonzerol=0 ]
again , @xmath467 .
if @xmath142 contains @xmath520 where @xmath185 is even , @xmath198 is odd , @xmath21 is an integer and @xmath521 , then @xmath142 commutes with an element of @xmath2 containing @xmath522 , which commutes with an element of @xmath2 containing @xmath523 , which commutes with an element of @xmath2 containing @xmath524 , which finally commutes with an element @xmath165 of @xmath2 containing @xmath525 , such that the transposition of @xmath165 is the same as the transposition of @xmath142 , namely not @xmath526 . by proposition [ mnonzerol=0 ] ,
hence @xmath510 .
the final case to consider is where @xmath14 is contained in the transposition of @xmath142 and @xmath327 is in an odd negative 1-cycle .
so @xmath142 contains @xmath527 for some @xmath196 and integers @xmath528 with @xmath198 odd .
if @xmath529 then subject to appropriate conjugation we can set @xmath530 . using proposition [ mnonzerol=0 ] in @xmath531
we see that @xmath142 is distance at most @xmath532 from the element @xmath165 where @xmath533 .
then @xmath165 commutes with @xmath534 , which commutes with @xmath535 which is distance 2 from @xmath42 .
thus @xmath510 .
the last case is where without loss of generality @xmath536 and we can assume @xmath470 is 2 or 5 . let @xmath537 be the other element of @xmath538 . then @xmath142 is distance 2 from some @xmath479 in @xmath2 containing @xmath539 where @xmath540 is determined by @xmath142 but we may choose @xmath164 arbitrarily , and @xmath541 are integers with @xmath199 odd . now let @xmath542
. then @xmath543 and @xmath544 .
what is @xmath545 ?
if @xmath546 then @xmath547 . set @xmath548 . now working in @xmath549 we see from proposition [ mnonzerol=0 ] that @xmath550 . if @xmath551 then @xmath552 . this time set @xmath553 .
then @xmath479 commutes with @xmath554 for some @xmath555 , which commutes with @xmath556 which commutes with @xmath165 , so @xmath557 .
hence in all cases @xmath510 , which completes the proof of lemma [ kois2 ] .
[ connectm=1l=0 ] suppose @xmath558 , @xmath462 , @xmath31 and @xmath24 and @xmath26 are both at least 2
. then @xmath4 is connected with diameter at most @xmath1 . if @xmath496 , then @xmath559 .
assume that @xmath560 .
suppose first that @xmath37 and let @xmath371 .
we use induction on @xmath26 to show that if the transposition of @xmath142 is not @xmath561 , then @xmath510 .
otherwise @xmath562 .
if @xmath502 the result holds by lemma [ kois2 ] .
if @xmath563 and the transposition of @xmath142 is not @xmath561 , then @xmath142 contains cycles @xmath564
where @xmath471 , @xmath565 , @xmath566 is a cycle of @xmath42 and @xmath567 and @xmath21 are all congruent modulo 2 .
now @xmath142 commutes with some @xmath517 containing the cycles @xmath568 for appropriate @xmath569 .
if we ignore @xmath566 and work in @xmath570 , then inductively @xmath571 .
hence @xmath510 .
suppose the transposition of @xmath142 is @xmath561 .
then @xmath142 commutes with some @xmath165 in @xmath2 that does not contain this transposition , and we have seen that @xmath572 . hence @xmath562 .
this completes the case @xmath37 .
+ if @xmath463 and @xmath502 then it is easy to see that @xmath142 is distance at most 2 from an element of @xmath2 containing the transposition @xmath573 .
thus @xmath142 is distance at most 3 from an element @xmath165 of @xmath2 containing @xmath574 . ignoring these 1-cycles we may work in @xmath492 and apply proposition [ mnonzerol=0 ] to see that @xmath575 .
hence @xmath467 . +
now suppose @xmath576 and @xmath563 .
if @xmath142 has a transposition containing an element @xmath470 with @xmath498 , then @xmath142 commutes with an element @xmath165 containing @xmath373 or @xmath577 ( choose whichever of these is a cycle of @xmath42 ) .
then we can ignore this cycle and work in @xmath578 .
inductively , using the base case @xmath502 , we see that @xmath579 .
. finally we deal with the case that every transposition of @xmath142 is of the form @xmath580 where @xmath581 . because @xmath582 , it must be the case that @xmath142 contains cycles : @xmath583 where @xmath584 , @xmath585 , @xmath586 , @xmath587 and @xmath588 .
then @xmath142 is distance 2 from an element @xmath165 with the cycles @xmath589 and @xmath590 .
now , working inductively in @xmath591 we see that @xmath474 .
hence @xmath467 .
[ n=5 ] if @xmath592 , @xmath37 , @xmath315 and @xmath477 , then @xmath593 .
let @xmath371 .
if the transposition is @xmath594 then using table [ tableg3 ] for @xmath595 we see that @xmath596 .
suppose the transposition of @xmath142 is @xmath597 where @xmath598 , and let @xmath599 be the remaining element of @xmath600 . then @xmath142 commutes with @xmath601 for some even integers @xmath185 and @xmath602 .
now @xmath479 commutes with @xmath603 , which commutes with @xmath42 .
so @xmath596 .
the remaining cases are ( interchanging 1 and 2 if necessary ) when @xmath604 or @xmath605 for appropriate @xmath606 and @xmath602 .
the following is a path of length at most 5 from either of these to @xmath42:@xmath607 @xmath142 , @xmath608 , @xmath609 , @xmath610 , @xmath611 , @xmath42 .
hence in all cases @xmath612 , which completes the proof .
we observe , because we will need it for lemma [ n=6 ] later , that the proof of lemma [ n=5 ] shows that @xmath613 in all cases except where ( modulo interchanging 1 and 2 , or 3 and 4 ) the transposition of @xmath142 is @xmath614 .
[ m>0l>0 ] suppose @xmath462 , @xmath615 and @xmath616 .
then @xmath302 is connected with diameter at most @xmath14 .
suppose @xmath14 is minimal such that @xmath302 is a counterexample , and let @xmath371 such that @xmath617 .
by lemma [ n=5 ] we can assume @xmath469 .
if @xmath449 then @xmath371 commutes with some @xmath472 containing @xmath618 . ignoring this 1-cycle we can work in @xmath619 to find a path to @xmath42 , which inductively is of length at most @xmath327 , which implies @xmath467 , contrary to our choice of @xmath142 . hence @xmath432
+ if elements @xmath470 and @xmath421 lying between @xmath620 and @xmath621 are contained in transpositions of @xmath142 , then @xmath142 commutes with some @xmath517 having the transposition @xmath597 for some @xmath185 ( if @xmath37 then we can set @xmath622 ) .
if we ignore this transposition of @xmath479 we can work in @xmath623 , which is either the case @xmath315 with a smaller @xmath23 , so inductively the graph has diameter at most @xmath326 , or ( if @xmath37 ) we can use theorem [ disc ] , in which case the graph has diameter @xmath327 . in either case , we see that @xmath142 is distance at most @xmath327 from the element @xmath99 of @xmath2 whose cycles are the same as @xmath42 except that @xmath99 has @xmath624 instead of @xmath625 .
since @xmath99 commutes with @xmath42 we have @xmath467 .
the same argument holds if @xmath626 .
if @xmath627 then similar reasoning shows again @xmath142 is distance at most @xmath327 from an element @xmath99 containing @xmath628 that commutes with @xmath42 .
so @xmath467 . thus none of these pairs @xmath196 exist in @xmath142 .
this implies @xmath629 .
moreover if @xmath630 then @xmath631 and without loss of generality the transpositions of @xmath142 contain @xmath632 and @xmath14 .
+ if there is some @xmath421 in a transposition of @xmath142 with @xmath633 and if @xmath634 , then @xmath142 commutes with some @xmath472 containing @xmath375 .
inductively we can work in @xmath635 to see that @xmath579 .
hence @xmath467 . similarly , as long as either @xmath497 or @xmath634 ( or both ) , if there is some @xmath470 in a transposition of @xmath142 with @xmath636 then inductively @xmath467 .
this means that if @xmath630 then @xmath477 and @xmath309 .
+ suppose that @xmath630 , @xmath477 and @xmath309 , so that @xmath637 .
we have observed that the transpositions of @xmath142 must contain @xmath638 and @xmath639 .
if 6 is not contained in an even negative 1-cycle of @xmath142 , then @xmath142 commutes with some @xmath165 containing a transposition @xmath597 for some @xmath185 , where @xmath640 . as at the start of this proof ,
more explicitly , inductively @xmath165 is distance at most 6 from an element @xmath99 containing @xmath642 that commutes with @xmath42 .
hence @xmath643 , a contradiction . thus @xmath37 .
+ if @xmath142 contains @xmath644 where @xmath645 then @xmath142 commutes with some @xmath517 containing @xmath577 .
so , using the result for @xmath646 , we get @xmath647 .
thus @xmath467 .
the same reasoning holds if @xmath497 and @xmath142 contains @xmath644 for some @xmath470 where @xmath636 .
suppose first that @xmath648 . since @xmath469 we must have @xmath634 or @xmath350 and @xmath503 .
thus the transposition of @xmath142 can not contain any @xmath470 with @xmath649 . hence without loss of generality @xmath142 contains @xmath650 .
suppose that @xmath142 contains @xmath651 where @xmath652 mod @xmath653 and either @xmath654 or @xmath655 .
then , @xmath142 commutes with some @xmath472 where @xmath165 contains @xmath656 for some @xmath185
. now @xmath42 commutes with some element @xmath429 containing @xmath657 . using theorem [ disc ] on @xmath658 ( that is , removing @xmath659 , from @xmath165 and @xmath429 )
we see that @xmath660 .
therefore , @xmath467 .
finally , we are reduced to the possibility that @xmath142 contains @xmath661 and any pair @xmath662 where @xmath663 and @xmath664 satisfies @xmath665 . since @xmath286 and @xmath648 , the only way this can occur is when @xmath666 and @xmath552 .
conjugating by an element of the centraliser of @xmath42 if necessary , we can assume that @xmath667 where @xmath21 and @xmath555 are even , and @xmath198 , @xmath330 are odd .
the following is a path from @xmath142 to @xmath42 in the graph : @xmath142 , @xmath668 , @xmath669 , @xmath670 , @xmath671 , @xmath672 , @xmath42 .
so @xmath673 , another contradiction .
+ the final case to consider is where @xmath37 and @xmath309 .
assume first that @xmath674 .
suppose that @xmath142 contains @xmath675 with @xmath567 both even .
then @xmath142 commutes with some @xmath472 containing @xmath676 for some @xmath199 . as observed at the start of this proof , @xmath579 .
hence @xmath467 .
now suppose @xmath142 does not contain @xmath677 .
then there is a set @xmath678 containing 1 , 2 and @xmath327 such that @xmath679 where @xmath185 and @xmath602 are both even and @xmath198 is odd . by replacing @xmath142 with a conjugate under the centraliser of @xmath42
, we can further assume that @xmath680 for some @xmath421 . let @xmath681 .
now , using the case @xmath682 on @xmath683 we see that @xmath684 . next we apply theorem [ disc ] to @xmath685 and @xmath686 ( noting that here there are no odd negative 1-cycles ) to see that @xmath687 .
hence @xmath688 , another contradiction . +
the remaining possibility is that @xmath38 , @xmath37 , @xmath689 .
but lemma [ n=6 ] immediately after this proof shows that this graph has diameter 6 , which is the final contradiction completing the proof of theorem [ m>0l>0 ]
. [ n=6 ] if @xmath38 , @xmath37 , @xmath477 and @xmath309 then @xmath690 .
we have @xmath691 .
let @xmath371 .
suppose @xmath142 contains @xmath692 for some @xmath21 . by theorem [ disc ] , the graph of @xmath693 in @xmath694 has diameter 5
. hence @xmath612 .
if @xmath142 contains @xmath695 where @xmath567 are even , then @xmath142 commutes with some @xmath479 containing @xmath692 for some @xmath21 , and we have just seen that @xmath696 . hence @xmath673 .
+ we next consider the cases where 1 and 2 are in different types of 1-cycle of @xmath142 . because interchanging 1 and 2 does not affect the distance of @xmath142 from @xmath42 , there are just three cases to consider here : @xmath142 contains @xmath697 , @xmath698 or @xmath699 , where @xmath185 is even and @xmath198 is odd .
in the first case we have @xmath700 where @xmath701 , @xmath567 are even and @xmath198 is odd . here
@xmath142 commutes with @xmath702 , which commutes with the element @xmath99 given by @xmath703 . glancing at table [ tableg4 ]
we observe that in @xmath694 elements at distance 4 or 5 from @xmath693 require at least one even label to be nonzero .
so @xmath704 .
hence @xmath612 .
for the second case , where @xmath142 contains @xmath698 , note that @xmath142 commutes with some @xmath479 containing @xmath697 .
hence , using the first case we get @xmath673 .
the third case is where @xmath142 contains @xmath699 . here
, @xmath142 commutes with some @xmath479 of the form @xmath705 .
the following is a path from @xmath479 to @xmath42 : @xmath706 so @xmath673 .
+ it remains to take care of the possibility that exactly one of 1 and 2 lies in the transposition of @xmath142 . here , without loss of generality , we can assume @xmath142 contains @xmath707 for @xmath471 .
if @xmath142 contains @xmath708 for @xmath198 odd , or @xmath709 , then @xmath142 commutes with some @xmath479 containing @xmath697 , for even @xmath185 , and as shown earlier in this proof , @xmath696 .
hence @xmath673 .
we may thus assume that @xmath710 . if @xmath711 then @xmath142 commutes with @xmath479 given by @xmath712 . by theorem [ n=5 ]
applied to @xmath713 we see that @xmath696 and so @xmath673 . so we can assume that @xmath714 . now @xmath142 is distance 4 from the element @xmath99 given by @xmath715 , as shown by the following path .
@xmath716 if @xmath717 , then @xmath718 so @xmath612 . if @xmath719 , then @xmath720 , so @xmath673 . if @xmath721 , then @xmath720 so @xmath673 .
since interchanging 3 and 4 does not affect the distance from @xmath42 , there is only one case left to deal with : @xmath722 and @xmath723 .
so @xmath724 . but
@xmath142 is the same distance from @xmath42 as @xmath725 , and @xmath165 commutes with @xmath429 given by @xmath726 for appropriate @xmath727 and @xmath555 . by the ` @xmath717 '
case above , @xmath728 . therefore @xmath729 .
this completes the proof .
[ [ proof - of - theorems - summary - and - main ] ] proof of theorems [ summary ] and [ main ] + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + suppose @xmath30 .
if @xmath31 then @xmath4 is disconnected by theorem [ mainhalf](i ) .
otherwise @xmath730 by theorem [ disc ] .
suppose @xmath37 and @xmath31 .
if one of @xmath24 and @xmath26 is 1 or the largest of @xmath24 and @xmath26 is 2 , then @xmath4 is disconnected by theorem [ mainhalf ] ( ii ) , ( iv ) and ( v ) .
otherwise @xmath731 by propositions [ m>0k=0 ] , [ mnonzerol=0 ] , lemma [ kois2 ] and theorem [ connectm=1l=0 ] .
suppose @xmath576 and @xmath31 .
if either of @xmath24 or @xmath26 is 1 , then @xmath4 is disconnected by theorem [ mainhalf](ii ) .
otherwise @xmath559 by propositions [ m>0k=0 ] , [ mnonzerol=0 ] and theorem [ connectm=1l=0 ] .
finally suppose @xmath462 and @xmath732 .
if @xmath733 then @xmath4 is disconnected by theorem [ mainhalf](iii ) .
otherwise @xmath559 by lemma [ n=5 ] , theorem [ m>0l>0 ] and lemma [ n=6 ] .
we first summarise the information on commuting involution graphs for @xmath734 , @xmath735 and @xmath736 . for fixed @xmath17 ,
the @xmath737 disc @xmath738 is the set of elements of @xmath2 which are distance @xmath62 from @xmath42 . since a length preserving graph automorphism interchanges classes with @xmath560 and those with @xmath739 , we list here only those classes with @xmath560 . in describing the elements we omit positive 1-cycles . in @xmath734
there are two connected graphs , both of diameter 2 . if @xmath740 then @xmath741 with the remaining elements of @xmath2 comprising the second disc .
if @xmath742 then @xmath743 , with all other elements of @xmath2 lying in the second disc
. tables [ tableg3 ] and [ tableg4 ] give , for each connected graph and each disc , a list of orbit representatives under the action of @xmath744 .
we use @xmath567 and so on to represent arbitrary even numbers , with @xmath745 $ ] being an arbitrary nonzero even number .
we use @xmath746 and so on for arbitrary odd numbers , with @xmath747 $ ] being any odd number other than 1 .
finally @xmath21 , @xmath555 and so on will be arbitrary integers with @xmath748 $ ] an arbitrary nonzero integer .
figure [ l0 ] is the collapsed adjacency graph for @xmath750 in @xmath751 , which has diameter 5 . in the graph
@xmath21 is an arbitrary integer , @xmath748 $ ] is an arbitrary non - zero integer and @xmath567 and @xmath727 are arbitrary even integers .
for simplicity we have only included shortest paths that is , we have omitted edges between nodes in the same disc .
99 f. ali , m. salman , and s. huang .
_ on the commuting graph of dihedral group
_ , communications in algebra vol .
44 , iss . 6 ( 2016 ) 23892401 .
a. nawawi and p. rowley .
_ on commuting graphs for elements of order 3 in symmetric groups _ , electron
. j. combin . 22
( 2015 ) , no .
1 , paper 1.21 , 12 pp . b. fischer .
_ finite group generated by 3-transpostions _ , i.invent .
math.13 ( 1971 ) , 232 - 246 . c. bates , d. bundy , s. perkins and p. rowley .
_ commuting involution graphs for symmetric groups _ , j. algebra , 266(1)(2003 ) , 133 - 153 .
c. bates , d. bundy , s. perkins and p. rowley . _ commuting involution graphs for finite coxeter groups _ , j. group theory 6(2003),461 - 476 .
m. giudici and a. pope .
_ on bounding the diameter of the commuting graph of a group _
, j. group theory , 17 ( 1)(2014 ) , 131 - 149 .
_ reflection groups and coxeter groups _ , cambridge studies in advanced mathematics , 29 ( 1990 ) .
s. perkins .
_ commuting involution graphs in the affine weyl group @xmath8 _ , arch . math .
86 ( 2006 ) , no . 1,16 - 25 .
_ conjugacy classes of involutions in coxeter groups _ , bull .
* 26 * ( 1982 ) , 115 .
y. segev and g.m .
_ anisotropic groups of type @xmath752 and the commuting graph of finite simple groups _ , pacific j. math .
202(2002 ) , 125226 . | in this article we consider commuting graphs of involution conjugacy classes in the affine weyl group of type @xmath0 .
we show that where the graph is connected the diameter is at most @xmath1 .
+ msc(2000 ) : 20f55 , 05c25 , 20d60 . |
usually , when the heterostructure consisting of coupled ferromagnetic ( fm ) and antiferromagnetic ( af ) layers is cooled in field below the neel temperature of its af component , it shows the asymmetric magnetization @xcite , which is referred to as the exchange bias effect .
furthermore , the exchange bias field , defined as the average of the two coercive fields , is observed to decrease with increasing the number of the consecutive field cycling , which is referred to as the training effect@xcite .
the exchange bias and training effects are very interesting and could be used in future spintronics@xcite and data storage .
usually , the fm layer is taken as a whole and the af layer consists of many grains .
the af grain is small enough to consists of a single domain , and some uncompensated domains ( or grains ) may be formed by defects or impurities@xcite and couple with each other and with the fm domains . as the heterostructure is cooled to a low temperature ,
the uncompensated spins in the grains and domains become locked - in and prefer to a unidirection in the interface , thus contribute to the magnetization shift@xcite . moreover , under the reversal of fm domains , the uncompensated grains or domains will be irreversibly reorganized@xcite and thus cause the training effect . the idea of domain states
was corroborated in some monte carlo simulations@xcite .
on the other hand , hoffmann@xcite considered the biaxial anisotropy of the af sublattices and solved it by variational method .
actual nonequilibrium dynamical properties of the magnetization are still waiting to be elucidated .
it is highly desirable and needed to systematically investigate the two effects in a unified theory . in this article
we use a unified monte carlo dynamical approach@xcite to study the fm / af heterostructure in order to investigate the exchange bias and training effect .
our simulated result shows the obvious shift of hysteresis loops and the cycling dependence of exchange bias .
the magnetization of uncompensated af layer is still open after the field cycling is finished .
the exchange bias fields decrease with decreasing the cooling rate or increasing the temperature and the number of the field cycling . with the simulations
, we shows the exchange bias can be manipulated by controlling the cooling rate , the distributive width of the anisotropy energy , or the magnetic coupling constants .
essentially , these two effects can be explained on the basis of the microscopically irreversible reversal of the af domains .
more detailed results will be presented in the following .
the remaining part of this paper is organized as follows . in next section
we shall define our model and discuss our simulation method . in section iii we shall present our simulated results and analysis . in section iv we shall discuss the microscopic mechanism for the phenomena in a unified way .
finally , we shall give our conclusion in section v.
according to experimental observations@xcite , for both compensated and uncompensated af layers the easy axis tends to form along external cooling field direction rather than later rotating field direction . in our model
the af layer consists of many af domains , and the fm layer consists of one single domain . assuming the cooling field is applied parallel to the af / fm interface , then all the easy axes of af and fm domains lie in the plane of the interface .
the coupled bilayers of af and fm domains are shown in the inset of fig .
the rectangles of the white pattern represent the af domains and the larger rectangle is the single fm domain
. the af domains couple to each other antiferromagnetically and the single fm domain couples to all the af domains ferromagnetically .
we define the @xmath0 axis along the common easy axis which lies in the interface plane .
we apply the external field to saturate the magnetization of the fm layer along the @xmath0 axis . for simplicity , we consider all the uncompensated spins in the af domains are the same .
we use @xmath1 to denote the spin vector of the @xmath2th af domain and @xmath3 to denote that of the single fm domain , where @xmath4 and @xmath5 are the uncompensated spin values and fm spin respectively .
then we write the hamiltonian of the bilayers in an external field as @xmath6 where @xmath7 and @xmath8 .
the first and second terms represents the anisotropy of the fm domain and the af ones , and @xmath9 and @xmath10 are the corresponding anisotropy constants .
the third term represents the zeeman energy of the moments due to the applied external field .
the fourth term represents the antiferromagnetic coupling among the af domains .
the last term represents the ferromagnetic coupling between the fm and af domains . using @xmath11 and @xmath12 to describe the angles of the @xmath2-th af moment and the fm moment deviating from the common easy axis
, we can express the energies of the fm domain and the @xmath2-th af as @xmath13 and @xmath14 where both @xmath15 and @xmath16 are the scalars taking either 1 or -1 .
thus for the @xmath2-th af domain the energy increment is @xmath17 , where @xmath18 , and for the fm domain the energy increment is @xmath19 , where @xmath20 .
we can express @xmath21 and @xmath22 as@xcite @xmath23\ ] ] and @xmath24\ ] ] as a result , to reverse its moment , the the fm layer must overcomes a barrier @xmath25 if @xmath26 , or @xmath27 if @xmath28 ; and the @xmath2-th af grain a barrier @xmath29 if @xmath30 , or @xmath31 if @xmath32 . if the condition @xmath33 or @xmath34 is satisfied , there is no barrier for the reversal .
actually , for the distribution of the af anisotropy energy we use a gauss function , @xmath35 $ ] , whose @xmath36 and @xmath37 are set to 30.0 mev and 50.0 mev unless stated otherwise
. the anisotropy energy of the fm domain is set 200.0 mev without losing main physics .
thus the reversal rate for a spin to reverse is @xmath38 , where @xmath39 is the energy barrier and @xmath40 is the characteristic frequency . in our simulations , @xmath40 is set to @xmath41/s .
we adopt a square lattice for the af domains and use @xmath42 as its size .
since we are only interested in the exchange bias and training effect at the nanoscale , the af lattice is enough to capture main physics .
furthermore , we assume the af domains have uniform moment 4.0 @xmath43 and the fm domain 2000 @xmath43 .
the coupling constant @xmath44 is set to 4.0 mev , and @xmath45 8.0 mev . in our simulations
the system is quenched from a high - enough temperature such as 610 k , at which the af layer is paramagnetic , to a low - enough temperature such as 10 k. the magnetization and exchange bias fields are calculated at the low temperature 10 k unless the temperature is explicitly stated otherwise .
the basic rate of changing temperature is @xmath46= 50 k / s .
the field sweeping rate is set to 0.5 t / s with the basic increment 0.1 t for each simulation step .
at first , we let the af / fm bilayers relax under a magnetic field of 5.0 t at a high temperature 610 k. this temperature is enough to make both the fm layer and the af layer remain paramagnetic .
when the temperature decreases , the average magnetization values of the two layers increases .
the external field makes the average magnetization of the fm layer have a large increase below 600 k , and reach nearly to the saturated value at 500 k. when the temperature becomes lower than 60 k , the average magnetization of the af layer looks like that of an antiferromagnet under an applied field and is dependent on the cooling rate @xmath47
. then , we further cool the bilayers under the same field . after the temperature reaches down to 10 k
, we start to change the field while keeping the temperature unchanged .
the field decreases from 5.0 t to -10.0 t and then increases back to 5.0 t for the first hysteresis . repeating the field cycling , we will make the second hysteresis loop .
the simulated results are shown in fig .
1 . the first two hysteresis loops of the fm ( a ) and af ( b ) layer at 10 k. the inset in ( a ) shows the af / fm bilayers . the hysteresis loop is obtained by changing the field in the order of p1-p2-p3-p4-p1 .
, width=302 ] as shown in fig .
1(a ) , the origin of the first hysteresis is clearly shifted in the negative field direction and shows the exchange bias .
the exchange bias field is defined as @xmath48 , where @xmath49 and @xmath50 is the coercivity of the left and right branches .
the left branch of the second hysteresis moves towards the positive direction , but the right branches of the first two loops almost coincide with each other . actually , any further loop almost does no difference in the right branch from the second hysteresis .
the shift of the second loop clearly demonstrates that the bilayers magnetization depends on the cycling history , which is known as training effect .
1(b ) shows the magnetization of the af layer , which drops largely and opens widely due to the irreversible reversal of the af domains after the first field cycling is finished .
the subsequent magnetization is more smooth but still not closed , indicating the continuing cycling dependence of exchange bias .
this is consistent with other monte carlo simulations@xcite .
temperature dependence of the exchange bias fields for the first two loops .
the exchange bias field is calculated at a given temperature after the system is cooled from 610 k to the temperature value .
the lines are the fitting curves in terms of the simple function defined in eq .
( 6).,width=302 ] we study the effect of the temperature @xmath51 on the exchange bias field , @xmath52 , for different loops .
our simulated exchange bias fields as functions of @xmath51 for the first two loops are shown in fig .
2 . for both of the two curves ,
the data can be fitted by the simple function @xmath53 where @xmath54 , @xmath55 , and @xmath56 are fitting parameters . for the fitting in fig .
2 , the parameters @xmath54 , @xmath55 , and @xmath56 takes 4.26 t , 17.95 k , and 1.58 for the first loop , and 3.94 t , 16.66 k , and 1.59 for the second loop . our results are consistent with experimental observation that the exchange bias field decreases with increasing temperature @xcite . the loop - number dependence of the exchange bias fields for the gaussian width @xmath36=30.0 and 20.0 mev .
the temperature is 10 k. all the data except for @xmath57 can be well fitted by a simple function @xmath58.,width=302 ] the exchange bias field is dependent on the field cycling number @xmath59 .
our simulated result from @xmath59=1 to @xmath59=9 is shown in fig .
3 . here , the temperature is kept at 10 k , and @xmath36 is set to 20.0 and 30.0 mev . for both of the curves in fig .
3 , the data points excepts of @xmath57 are well fitted by the simple function @xmath58 , where @xmath60 , @xmath61 , and @xmath62 are the fitting parameters , taking 0.23 t , 2.26 t , and 0.76 for @xmath63 mev , and 0.22 t , 1.51 t , and 0.74 for @xmath64 mev .
the value of @xmath65 usually is substantially above the extrapolation of the other @xmath66 ( @xmath67 ) .
these simulated results are in good agreement with experimental observation@xcite .
the cooling - rate dependence of the exchange bias fields for the first two loops .
the lines are the fitting curves in terms of eq .
, width=302 ] since the training effect reflects the non - equilibrium dynamical magnetization which is caused by the irreversible reversal of meta - stable domains formed during quenching , the quenching rate must affect the exchange bias field . by changing the cooling rate @xmath47 , we study the exchange bias field as the function of quenching rate @xmath47 . the result is shown in fig .
4 . for both of the loops ,
the data be well fitted by @xmath68 where @xmath69 and @xmath70 are 0.292 t and 20847 for the first loop , and 0.262 t and 16248 for the second loop .
the exchange bias field at 10 k increases logarithmically with increasing the quenching rate .
the exchange bias fields as functions of the coupling - constants @xmath44 ( a ) and @xmath45 ( b ) for the first two loops . ,
width=302 ] the exchange bias field as a function of the gaussian width @xmath36 for the first two loops . , width=302 ] it is interesting to investigate the dependence of the exchange bias field on the coupling constants @xmath44 and @xmath45 . the simulated results are shown in fig .
5 . as shown in fig .
5(a ) , the exchange bias field decreases with increasing @xmath44 .
the training effect is nearly unchanged when @xmath44 changes from 0 to 2mev , but diminishes to zero quickly with increasing @xmath44 from 2mev .
when @xmath44 is larger than 6 mev , the exchange bias field already becomes very small and the training effect is actually zero . in contrast , the @xmath45 data points in fig .
5(b ) can be well fitted by the simple function @xmath71 , where @xmath72 and @xmath73 are the fitting parameters , taking 0.44 t and 4.01 mev for the first loop , and 0.51 t and 4.49 mev for the second loop .
it is clearly shown that the exchange bias field increases exponentially as @xmath45 increases .
both of the the exchange bias field and the training effect can be enhanced by decreasing @xmath44 and increasing @xmath45 , or by increasing @xmath74 .
this is consistent with experimental trend@xcite . in fig .
6 we shows how the distributive width @xmath36 of the af anisotropy affects the exchange bias fields . clearly the exchange bias field increases with @xmath36 , and so does the training effect . experimentally , the width can be increased by the additional nonmagnetic impurities and the enhanced roughness of the af crystalline phases .
this implies that the rougher the af crystalline phases are , the larger the exchange bias and training effect .
our result reveals that the exchange bias field is determined by both the coupling constants and the distributive width of the af domain anisotropy .
these are useful to completely understand the phenomenon@xcite .
after being cooled down to the low temperature , the af layer has a non - zero net fm moment @xmath75 due to the driving of both the field and the fm layer .
assuming there are @xmath76 af domains , on average we have the moments in part of all the @xmath76 af domains aligning parallel although they are coupled with af interactions .
the exchange bias field is determined by the effective moment @xmath75 , the difference of @xmath75 between the first two loops determines the training effect . naturally , both the exchange bias field and the training effect increase with increasing @xmath45 and with decreasing @xmath44 , as shown in fig .
5 . actually , small @xmath44 does not affect the effects , but larger @xmath44 than 2mev is harmful to the effects at 10 k. in addition , it is easily understood that @xmath75 decreases with increasing the temperature @xmath51 . as a result , both the exchange bias field and the training effect decrease with increasing @xmath51 , as shown in fig .
2 . the exponential description in eq . ( 6 ) reflects the fact that moment reversals are thermally activated .
it is reasonable that both the exchange bias field and the training effect increase with increasing the cooling rate @xmath47 , as shown in fig .
this is mainly because the average magnetization of the af layer increases with @xmath47 when the temperature is below 50 k. when the cooling rate approaches to zero , both the exchange bias field and the training effect should be zero . in another word
, our results should approach to those of corresponding equilibrium systems when the cooling rate @xmath47 approaches to zero . as shown in fig .
6 , both the exchange bias field and the training effect are nearly zero when the distributive width @xmath36 of the af anisotropy energy is smaller than 15 mev , but they increase substantially with increasing @xmath36 for @xmath77 mev .
this means that the effects are dependent on a wide distribution of the anisotropy energy .
this can be understood in terms of the changing of the energy barriers with the external field . from p1 to p2 in fig .
1(a ) , the effective barrier of the fm layer decreases but is still high enough to avoid the reversal , but meanwhile , more and more spins of the af domains are reversed due to their lower energy barriers . at the point p2 , the fm moment is reversed with the help of the field and the reversing of the af domains .
anyway , some af domains with high energy barriers have their moments unchanged , even after the fm layer has been reversed , and thus there is a net average moment of the af domains parallel to the moment of the fm layer .
this net average moment increases with the distributive width @xmath36 .
this explains the increasing of the exchange bias field and training effect with increasing @xmath36 .
the more the field cycling loops , the longer the time . actually , this is similar to reducing the cooling rate @xmath47 in effect . as a result
, the exchange bias field decreases with increasing the number of the field cycling .
the turning point of the time scale causes the largest drop happens between the first loops .
in summary , we use a unified monte carlo dynamical approach@xcite to study the fm / af heterostructure in order to investigate the exchange bias and training effect . the magnetization of uncompensated af layer is still open after the first field cycling is finished .
our simulated result shows the obvious shift of hysteresis loops ( exchange bias ) and the cycling dependence of exchange bias ( training effect ) .
the exchange bias fields decrease with decreasing the cooling rate or increasing the temperature and the number of the field cycling . with the simulations ,
we show the exchange bias can be manipulated by controlling the cooling rate , the distributive width of the anisotropy energy , or the magnetic coupling constants .
essentially , these two effects can be explained on the basis of the microscopical coexistence of both reversible and irreversible moment reversals of the af domains .
our simulated results are useful to really understand the magnetization dynamics of such magnetic heterostructures which should be important for spintronic device and magnetic recording media @xcite .
this unified nonequilibrium dynamical method should be applicable to other exchange bias systems .
this work is supported by nature science foundation of china ( grant nos .
10874232 , 10774180 , and 60621091 ) , by the chinese academy of sciences ( grant no .
kjcx2.yw.w09-5 ) , and by chinese department of science and technology ( grant no .
2005cb623602 ) .
s. bruck , g. schutz , e. goering , x. s. ji , and k. m. krishnan , phys .
lett . 101 , 126402 ( 2008 ) .
m. gruyters and d. schmitz , phys .
100 , 077205 ( 2008 ) .
y. ijiri , t. c. schulthess , j. a. borchers , p. j. van der zaag , and r. w. erwin , phys .
99 , 147201 ( 2007 ) .
s. brems , d. buntinx , k. temst , and c. v. haesendonck , phys .
95 , 157202 ( 2005 ) .
j. nogues and i. k. schuller , j. magn .
192 , 203 ( 1999 ) .
a. hochstrat , c. binek , and w. kleemann , phys .
b 66 , 092409 ( 2002 ) .
t. zhao , a. scholl , f. zavaliche , k. lee , m. barry , a. doran , m. p. cruz , y. h. chu , c. ederer , n. a. spaldin , r. r. das , d. m. kim , s. h. baek , c. b. eom , and r. ramesh , nat .
mater . 5 , 823 ( 2006 ) ; r. ramesh and n. a. spaldin , nat .
6 , 21 ( 2007 ) .
v. laukhin , v. skumryev , x. mart , d. hrabovsky , f. sanchez , m. v. cuenca , c. ferrater , m. varela , u. luders , j. f. bobo , and j. fontcuberta , phys .
97 , 227201 ( 2006 ) .
p. borisov , a. hochstrat , x. chen , w. kleemann , and c. binek , phys .
94 , 117203 ( 2005 ) .
p. miltenyi , m. gierlings , j. keller , b. beschoten , and g. guntherodt , phys .
84 , 4224 ( 2000 ) .
j. i. hong , t. leo , d. j. smith , and a. e. berkowitz , phys .
96 , 117204 ( 2006 ) .
f. u. hillebrecht , h. ohldag , n. b. weber , c. bethke , and u. mick , phys .
86 , 3419 ( 2001 ) .
m. bode , e. y. vedmedenko , k. v. bergmann , a. kubetzka , p. ferriani , s. heinze , and r. wiesendanger , nat .
5 , 477 ( 2006 ) .
w. kuch , l. i. chelaru , f. offi , j. wang , m. kotsugi , and j. kirschner , nat . mater . 5 , 128 ( 2006 ) . f. nolting , a. scholl , j. stohr , j. w. seo , j. fompeyrine , h. siegwart , j. p. locquet , s. anders , j. luning , e. e. fullerton , m. f. toney , m. r. scheinfein , and h. a. padmore , nature ( london ) 405 , 767 ( 2000 ) .
x. p. qiu , d. z. yang , s. m. zhou , r. chantrell , k. o. grady , u. nowak , j. du , x. j. bai , and l. sun , phys .
101 , 147207 ( 2008 ) .
t. hauet , j. a. borchers , p. mangin , y. henry , and s. mangin , phys .
96 , 067207 ( 2006 ) .
u. nowak , k. d. usadel , j. keller , p. miltenyi , b. beschoten , and g. guntherodt , phys .
b 66 , 014430 ( 2002 ) .
a. hoffmann , phys .
93 , 097203 ( 2004 ) .
y. li and b .-
liu , phys .
b 73 , 174418 ( 2006 ) ; phys .
96 , 217201 ( 2006 ) ; b .-
liu , k .- c .
zhang , and y. li , front .
china 2 , 424 ( 2007 ) .
w. zhu , l. seve , r. sears , b. sinkovic , and s. s. p. parkin , phys .
86 , 5389 ( 2001 ) .
m. gruyters , phys .
95 , 077204 ( 2005 ) . v. k. valev , m. gruyters , a. kirilyuk , and t. rasing , phys .
96 , 067206 ( 2006 ) .
j. camarero , j. sort , a. hoffmann , j. m. g. martin , b. dieny , r. miranda and j. nogues , phys .
95 , 057204 ( 2005 ) .
s. brems , k. temst , and c. v. haesendonck , phys .
99 , 067201 ( 2007 ) .
v. skumryev , s. stoyanov , y. zhang , g. hadjipanayis , g. dominique , and j. nogues , nature ( london ) .
423 , 850 ( 2003 ) . | we investigate the exchange bias and training effects in the fm / af heterostructures using a unified monte carlo dynamical approach .
this real dynamical method has been proved reliable and effective in simulating dynamical magnetization of nanoscale magnetic systems .
the magnetization of the uncompensated af layer is still open after the first field cycling is finished .
our simulated results show obvious shift of hysteresis loops ( exchange bias ) and cycling dependence of exchange bias ( training effect ) when the temperature is below 45 k. the exchange bias fields decrease with decreasing the cooling rate or increasing the temperature and the number of the field cycling . with the simulations
, we show the exchange bias can be manipulated by controlling the cooling rate , the distributive width of the anisotropy energy , or the magnetic coupling constants .
essentially , these two effects can be explained on the basis of the microscopical coexistence of both reversible and irreversible moment reversals of the af domains .
our simulated results are useful to really understand the magnetization dynamics of such magnetic heterostructures .
this unified nonequilibrium dynamical method should be applicable to other exchange bias systems . |
the vacuum energy produces measurable forces between neutral bodies @xcite . but even in the presence of a single body , the modes of the electromagnetic field are disturbed and produce a self - energy .
although irrelevant for the calculation of forces between different objects , this self - energy is in principle observable through its coupling to gravity . in the case of perfect conductors , the vacuum energy density , or more generally the energy - momentum tensor ,
diverges near the boundaries , as noted for the first time by deutsch and candelas a long time ago @xcite .
these divergences are not the usual ones in quantum field theory , because they are present in the _ already renormalized _ energy - momentum tensor .
the origin of the divergences is the unphysical assumption of perfect conductivity for all modes of the electromagnetic field since , on physical grounds , one expects any material to become transparent at high energies .
more generally , divergences in the renormalized energy - momentum - tensor are present even for non - perfect conductors , as long as there is a sharp boundary between two media with different electromagnetic properties . in this case
the origin of the divergences is that , for modes with extremely small wavelengths , it is unphysical to assume a sharp boundary , since the transition region between media becomes larger than the wavelength of the high frequency modes . the problem of the divergences in the self - energies has been considered in flat spacetime in a series of works by graham et al @xcite . it was pointed out that , if the boundary conditions are replaced by an interaction with a second field , the model is renormalizable and the self - energies are not divergent . in those works ,
the authors considered a toy model consisting of a vacuum scalar field @xmath0 coupled to a background ( classical field ) @xmath1 , with an interaction of the form @xmath2 .
the sharp limit corresponds to consider a discontinuous @xmath1 , and the perfect conductor " limit to take @xmath3 at a particular point , imposing dirichlet boundary conditions on the vacuum field at that point .
when these limits are taken , the self - energy depends on the ultraviolet cutoff , i.e. on the specific interaction between the high energy modes of the quantum field and the microscopic degrees of freedom of the body .
recently , milton @xcite and bouas et al @xcite considered a similar problem , computing the energy density for a scalar vacuum field in some particular potentials ( soft walls " ) , showing that the energy density near a potential barrier is finite for sufficiently smooth potentials . for
a discussion of some aspects of the coupling of the casimir energy to gravity see ref.@xcite .
a complete analysis of the casimir self - energies and their eventual gravitational implications must be performed in the context of quantum field theory in curved spacetimes .
following ref.@xcite , one could replace the boundary conditions by interactions with a classical field , and show not only that the matter sector of the theory is renormalizable , but also that the usual divergences in the energy - momentum tensor can be absorbed in the coupling constants of the gravitational sector , resulting in finite and well defined semiclassical einstein equations when the classical field is sufficiently smooth .
the divergences associated to the unphysical limits will reappear when considering discontinuous classical fields .
the aim of the present paper is to provide such analysis .
we will consider a scalar vacuum field in curved spacetimes , coupled to a classical field @xmath1 that models the mirror " .
we will show that , for smooth @xmath1 , the model is renormalizable using the standard renormalization procedure for quantum fields in curved spacetimes .
we will provide some examples to illustrate the appearance of divergences for sharp interfaces .
it is worth remarking that similar divergences in the renormalized energy - momentum tensor do appear even for free fields in curved spacetimes , if one assumes that the spacetime metric is not sufficiently smooth .
a well known example , in cosmology , is the divergence in the energy density that appears when one considers an abrupt transition between two different epochs ( e.g. de sitter space to radiation dominated universe @xcite ) , that involves a discontinuity in the curvature tensor .
a less known example is the divergence of the vacuum polarization around spherically symmetric objects , at a sharp interface between the object and the vacuum @xcite .
the paper is organized as follows . in the next section
we introduce the model . in section 3
we prove that the divergences in the energy - momentum tensor can be absorbed into the bare constants of the theory , yielding finite semiclassical einstein equations . in section 4
we present some explicit calculations for weakly coupled mirrors , showing that as long as the potential that models the mirror is smooth enough , there are no divergences in the renormalized energy - momentum tensor and in the renormalized coincidence limit of the two point function , @xmath4 . on the other hand , discontinuities in the potential or in its first two derivatives
produce infinite answers .
section 5 contains a discussion of the main results .
we will work with natural units @xmath5 and metric signature @xmath6 .
we consider a quantum vacuum field @xmath0 interacting with a background classical field @xmath1 on a curved spacetime .
the action of the complete system is @xmath7 where @xmath8 and @xmath9 , .
\label{eq : accion gravitatoria}\ ] ] here @xmath10 , @xmath11 , @xmath12 is the cosmological constant , and @xmath13 , i=1,2,3 are dimensionless parameters .
the terms quadratic in the curvature are needed in order to renormalize the theory , as is the self - interaction term for @xmath1 .
note that the classical field @xmath1 provides a position - dependent mass for the field @xmath0 , and therefore the propagation of @xmath0 will be suppressed in regions where the mass is very high . in this sense @xmath1 models a mirror " .
for example , a thin mirror located at @xmath14 is described by the interaction @xmath15 and the perfect conductor limit corresponds to @xmath16 .
the classical field equations are @xmath17 @xmath18 @xmath19 where @xmath20 and @xmath21 the tensors @xmath22 come from the variation of the terms quadratic in the curvature contained in the gravitational action . the classical energy - momentum tensor for the @xmath1 field is @xmath23\sigma^2\ , , \end{split } \label{eq : tuv2}\ ] ] while @xmath24 is the energy - momentum tensor for a free field with variable mass @xmath25 , that is , @xmath26\phi^2\ , .
\end{split } \label{eq : tuvphi}\ ] ] we consider now the semiclassical version of the theory , in which the field @xmath0 becomes a quantum field while @xmath1 and @xmath27 are treated classically . the heisenberg equation for the quantum operator associated to @xmath0
is given by the classical eq.([eq : phi ] ) .
the evolution equations for the classical backgrounds are obtained by taking the mean value of the classical eqs.([eq : sigma ] ) and ( [ eq : einstein ] ) : @xmath28 @xmath29 we have therefore established a well defined model to study the divergences of the self - energy in casimir calculations . the quantities @xmath30 and @xmath31 are formally divergent .
the divergences must be absorbed into the bare constants of the theory .
the renormalized version of @xmath31 contains information about the casimir effect ( force between different objects ) , as well as the self - energy .
the gravitational effects produced by the vacuum energy can in principle be computed by considering it as a source of the semiclassical einstein equations .
the theory of interacting fields in curved spacetimes can be renormalized using a precise covariant procedure @xcite . in the present model
it is necessary to introduce minor modifications to take into account that not only the metric but also one of the interacting fields is treated classically
. it will be particularly useful to adapt the renormalization method described in ref.@xcite .
it is shown there that to analyze the renormalizability of @xmath32 theory in curved spaces at the level of the equations of motion , one can split the field as @xmath33 , where @xmath34 is the mean value of the field and @xmath35 is the quantum operator that describes the fluctuations around the mean value . to one loop order , @xmath35 satisfies a free field equation with a variable mass .
the situation in our model is very similar , since the quantum field @xmath0 can be thought of a free field with variable mass .
therefore , we can follow closely ref.@xcite .
we define the renormalized quantities @xmath36 where @xmath37 and @xmath38 are constructed using the schwinger dewitt expansion up to fourth and second adiabatic order respectively @xcite .
the divergences present in these quantities are to be absorbed into the bare constants of the theory .
the usual schwinger dewitt expansion for the propagator of a massive field reads @xcite @xmath39 \end{split } \label{eq : sdw2}\ ] ] where @xmath40 is the feynman propagator , @xmath41 is the van - vleck determinant , @xmath42 is one half of the geodesic distance between @xmath43 and @xmath44 and @xmath45 is the number of spacetime dimensions , that acts as a regulator .
the functions @xmath46 are defined by a set of recursive equations that follows from imposing the equation for the propagator .
when the field has a variable mass , this expansion can be generalized to @xcite @xmath47\ , , \end{split } \label{eq : sdw}\ ] ] where @xmath48 $ ] with @xmath49 .
the quantities to be subtracted to cancel the divergences of @xmath50 and @xmath51 are @xcite @xmath52\ ] ] and @xmath53 _ { ; \mu \nu})_{ad4 } + \tfrac{1}{2 } ( \tfrac{1}{2 } - \xi_1 ) ( [ g_f^{sd}]_{;\mu \nu})_{ad4 } + \right.\\ & \left .
- \tfrac{1}{2 } ( \xi_1- \tfrac{1}{4})g_{\mu \nu } ( \box [ g_f^{sd}])_{ad4 } - \tfrac{1}{2}\xi_1 r_{\mu \nu}[g_f^{sd}]_{ad2 } \right\}\ , .
\end{split } \label{eq : tuv y g ad}\ ] ] we emphasize that we are not using a point - splitting regularization but a dimensional regularization
. we will substitute @xmath31 and @xmath30 in the semiclassical equations by @xmath54 and @xmath55 , respectively , and absorb the infinities contained in @xmath56 and @xmath38 into the bare constants appearing in those equations .
the coincidence limit of the two point function for a field with variable mass is given by @xcite @xmath57 where @xmath58 is an arbitrary constant with dimensions of mass .
inserting this expression into the semiclassical equation for @xmath1 and writing the bare constants in terms of renormalized ones @xmath59 we obtain @xmath60 \sigma + \frac{1}{3 ! } ( \lambda_{2r } + \delta \lambda_2 ) \sigma^3 + \frac{\lambda_{2r } \,\sigma}{32 \pi^2}\big[m^2 + \\ \\ & + ( \xi_1 - \frac{1}{6 } ) r \big ] \mbox{ln } \frac{m^2}{\mu ^2 } + \frac{\lambda_1 \sigma}{16 \pi^2 ( n-4 ) } \left(\frac{m^2}{\frac{n}{2}-1 } + ( \xi_1 - \frac{1}{6 } ) r \right ) + \frac{\lambda_1}{2 } \sigma \langle \phi^2\rangle _ { ren } = 0\ , .
\end{split } \label{eq : sigma3}\ ] ] therefore , the divergences can be absorbed with the counterterms @xmath61 where @xmath62 and @xmath63 are finite contributions ( they vanish in the minimal prescription scheme ) .
note that in the conformal case ( @xmath64 and @xmath65 ) , only a counterterm for the self - coupling of @xmath1 is needed .
let us now consider the renormalization of the gravitational sector of the theory .
as @xmath66 , the divergent part of @xmath56 is given by @xcite @xmath67-\frac{\lambda_1}{16 \pi^2 ( n-4 ) } \bigg[(\xi_1 - \tfrac{1}{6})(\sigma^2_{;\mu \nu } + g_{\mu \nu}\box \sigma^2 ) - \right .
\\ & - \left .
\sigma^2 \left ( - \frac{\xi_1 - \frac{1}{6}}{\frac{n}{2}-1 } g_{\mu \nu } + \frac{g_{\mu \nu}}{\frac{n}{2}(\frac{n}{2}-1 ) } ( m_1 ^ 2 + \tfrac{\lambda_1}{4}\sigma^2)\right)\right]\ , .
\end{split } \label{eq : tuv div}\ ] ] this expression contains both geometric divergences and divergences dependent on the classical field @xmath1 ( and its derivatives ) .
the former should be absorbed into a redefinition of the gravitational constants appearing in the left hand side of the semiclassical einstein equations , and the latter into a redefinition of the constants associated to the @xmath1 field , which appear in @xmath68 .
inserting eq.([eq : tuv div ] ) into eq.([eq : einstein2 ] ) one can show that the divergences dependent on @xmath1 can be absorbed using the same counterterms given in eq.([eq : redefiniciones1 ] ) .
this is a non trivial check of our calculations , and a necessary condition for the renormalizability of the theory . on the other hand ,
the geometric divergences can be absorbed into the gravitational constants , by choosing the following counterterns @xmath69 @xmath70 @xmath71 [ eq : redefiniciones2 ] @xmath72 @xmath73 [ eq : redefiniciones3 ] this completes the proof of the renormalizability of the model , and is one of the main results of this paper .
we have shown that , if the presence of a mirror is modeled by the interaction of the vacuum field with a classical background field , the divergences in the vacuum expectation value @xmath74 can be absorbed into the bare constants appearing in the semiclassical einstein equations .
not only the gravitational constants are renormalized but also the bare constants associated to the lagrangian of the classical background field .
the renormalizability is valid for one or more mirrors of arbitrary shape , as long as they can be described by a smooth function @xmath1 .
note that the divergences in the energy - momentum tensor are considerably simpler in the conformal case @xmath64 and @xmath65 . as a final remark
, we stress that the quantities @xmath75 and @xmath76 will be finite for a sufficiently smooth background field @xmath1 and spacetime metric @xmath27 .
we will illustrate this in the next section .
in this section we present simple expressions for evaluating explicitly @xmath77 and @xmath78 for a given background potential @xmath79 , within a weak field approximation . for simplicity we will work in minkowski space , and set @xmath64 so that the quantum field is massless .
the sense in which we define the weak field approximation is that we require @xmath80 , or in other words @xmath81 .
this can be achieved by having a weak coupling @xmath82 , or a weak and/or rapidly varying field @xmath79 .
the calculation will stay at the lowest order in @xmath82 .
the procedure we follow is similar to the one used in @xcite .
it is based in solving the equation for the feynman green function , @xmath83 to first order in @xmath82 , and using the resulting @xmath84 to compute @xmath85 and @xmath86 in the coincidence limit .
expanding eq.([gf ] ) by writing @xmath87 in a @xmath82 expansion and discarding the second - order term , we obtain @xmath88 this is solved explicitly using a momentum - space representation for @xmath89 : @xmath90 after switching variables to @xmath91 and @xmath92 , the @xmath93-integrals can be done by passing to @xmath45-dimensional space and employing standard dimensional regularization techniques , with the result in the coincidence limit being : @xmath94 here @xmath58 is an arbitrary mass scale introduced in the regularization procedure .
the divergence in the first term , which is purely local , can be absorbed in the coefficient @xmath95 of the action , as can be seen by comparison with the second line of eq.([eq : redefiniciones1 ] ) .
hence in accordance with eq.([phi2imgf ] ) we have @xmath96 for example , in the particular case in which the background potential depends only on the spatial coordinate @xmath97 , we shall have : @xmath98 this expression can be used to compute @xmath99 for the case of one or several parallel flat mirrors , which are `` almost transparent '' in the sense that the expression is valid to first order in the coupling @xmath82 .
we will prove now that according to this model @xmath99 is everywhere finite if the background potential @xmath100 is an integrable @xmath101 function . under this assumption
, it follows as a corollary of the riemann - lebesgue lemma that the fourier transform of @xmath100 ( which we note by @xmath102 ) falls off faster than @xmath103 .
hence @xmath104 is integrable , from which it follows that its fourier transform is in turn well - defined and finite at all points .
since according to ( [ fourierdoble ] ) it is precisely the fourier transform of @xmath105 that gives the nonlocal part of @xmath99 , and the local part is the term involving @xmath106 , it follows that both the local and the nonlocal parts of @xmath99 are finite for all @xmath97 if @xmath100 is an integrable @xmath101 function .
we have therefore proved that the divergences in @xmath50 are removed if the background field modeling the mirrors is sufficiently well - behaved .
we are not able to give a general proof using weaker assumptions than @xmath101 continuity .
however , we will see that in simple concrete examples the assumption of @xmath107 continuity is sufficient to obtain finite results .
let us first consider a discontinuous mirror of width @xmath108 , specified by the background potential : @xmath109 where @xmath110 is the unit step function .
the parameter @xmath111 has length dimension and is introduced so that @xmath100 is normalized by @xmath112 . introducing this expression into eq.([fourierdoble ] )
we obtain as a result : @xmath113\ ] ] where we omit the local contribution coming from the @xmath114 term .
notice that the result diverges logarithmically at the boundary of the mirror , where @xmath100 is discontinuous .
consider next a continuous mirror given by @xmath115 wherein the discontinuities in eq.([discont ] ) are resolved with linear interpolations of width @xmath116 . for this mirror
the vacuum polarization is computed to be : @xmath117,\end{aligned}\ ] ] where we have omitted again a local contribution proportional to @xmath118 .
this result is everywhere finite ; it can be compared to that of eq.([phi2dis ] ) in the plot in figure 1 . for a discontinuous potential ( red ) and a continuous one ( blue ) , with @xmath119 and @xmath120 .
note that @xmath121 diverges at the points where the background field @xmath1 is discontinuous , and that the divergences disappear when @xmath1 is continuous.,scaledwidth=80.0% ] we turn now to the study of @xmath78 to first order in @xmath82 . from the expression ( [ eq : tuvphi ] ) it follows that @xmath51 ( we drop the @xmath0 label from now on ) is obtained from the coincidence limit of @xmath122 as : @xmath123\bigg|_{x = x'}\,.\end{aligned}\ ] ] the coincidence limit of the last term is @xmath124 , and since @xmath50 is @xmath125 we may discard this term in our first - order approximation .
the coincidence limit of the remaining term can be performed in an analogous way to that of @xmath50 , by passing to @xmath45-dimensional space , using the momentum space representation ( [ gfpspace ] ) for the first - order @xmath122 , and using standard dimensional regularization techniques .
the divergences appearing are proportional to @xmath126 , and can be absorbed into a redefinition of @xmath95 as we did before .
after the renormalization is performed and we return to @xmath127 we are left with @xmath128 which is analogous to ( [ fourierdoble4d ] ) for @xmath129 . for the case in which @xmath118 depends only on @xmath97
we get by analogy with ( [ fourierdoble ] ) : @xmath130 comparing with our arguments showing finiteness for @xmath99 it is clear that the components of @xmath131 will not diverge as long as @xmath132 is integrable . to ensure this
it is sufficient to require @xmath100 to be an integrable @xmath133 function , since then its fourier transform falls off faster than @xmath134 .
however , for the simple particular example of a mirror similar to ( [ mirrorcont ] ) but with polynomials with higher degree of continuity interpolating between @xmath135 and @xmath136 , we have checked that @xmath101 continuity of @xmath100 is enough to render the result finite .
we stress at this point that the results obtained in this section have their gravitational counterparts : as shown in ref.@xcite , when one computes @xmath99 and @xmath131 for a free quantum field in a curved background , both quantities diverge at the points where the background is not sufficiently smooth .
the examples described in ref.@xcite refer to the vacuum polarization around spherically symmetric objects , and divergences show up for sharp interfaces , where the matter density is discontinuous .
the divergences disappear when the matter density and its first two derivatives are continuous across the interface .
similarly , in a cosmological context , divergences in the renormalized energy - momentum tensor are removed when one replaces an abrupt transition of the scale factor by a smoother transition in which the scale factor and its two first derivatives are continuous .
the concrete example of the transition between the inflationary period and radiation domination is discussed in detail in ref .
@xcite .
high vacuum energy densities in the presence of boundaries may potentially produce large gravitational effects . in our model , the usual divergences are rendered finite when the sharp boundary is replaced by a sufficiently smooth background field .
however , the quantum vacuum energy does attain high values if the background field varies over short distances .
we can make a crude estimation of the maximum value of the energy density using dimensional analysis . from eq.([fourierdobletmunu ] ) , it is not difficult to see that if @xmath100 is varying over distances of order @xmath116 then we should expect the maximum value of @xmath131 to be of order @xmath137 , where @xmath138 is a representative value attained by @xmath100 . the quantity @xmath139 has units of ( length)@xmath140 , and determines the properties of the mirror .
therefore it will be identified with the square of the plasma frequency @xmath141 . restoring @xmath142 and @xmath143 factors ,
the mass density associated to the vacuum fluctuations is of order @xmath144 .
if we take @xmath116 to be in the range of the @xmath145 meters , representing a smearing of the conductor s sharp boundary over atomic length scales , and @xmath141 to be in the range of @xmath146 hz , a typical value of the plasma frequency , then we see that firstly , the first - order approximation assumed in this section is validated , and secondly , the peak values of the quantum energy densities at the boundary of the conductor are of order @xmath147 , too small to be gravitationally detectable .
in this paper we analyzed the coupling of the vacuum self - energy to gravity .
we considered a model in which the presence of the bodies that disturb the modes of the quantum fields is described by the interaction with a background classical field .
we have shown that the divergences in the energy - momentum tensor of the quantum fields are consistent with the semiclassical einstein equations , that is , they can be absorbed into the bare constants of the theory .
as expected , the divergences in the renormalized energy - momentum tensor noticed in previous works only appear when considering unphysical limits of perfect conductivity and/or sharp interfaces , that in our model would correspond to take a non - smooth background field .
we have shown that modeling the mirrors in the casimir effect with a background potential @xmath118 removes the divergences attached to the boundaries , as long as @xmath118 is a sufficiently smooth integrable function ( @xmath101 continuity is sufficient for @xmath148 and @xmath133 continuity for @xmath149 , though weaker assumptions are enough in simple examples ) .
the proof is carried to first order in the coupling @xmath82 .
these results are in tune with those found in refs.@xcite , where finiteness of the results is shown with exact calculations for some particular examples of @xmath118 , and are analogous to those in refs.@xcite for the vacuum polarization of free fields in curved backgrounds , when the metric and its first two derivatives are continuous .
we expect that similar conditions for finiteness can be obtained beyond the perturbative approximation .
we have considered a toy model for a vacuum scalar field .
we have presented some specific examples to illustrate the appearance of divergences for sharp interfaces .
the examples suggest that the gravitational effects of the self - energy near mirrors are extremely small for realistic values of the relevant parameters .
we expect similar results to be valid for more realistic models involving the electromagnetic field .
this work was supported by uba , conicet and anpcyt .
p. w. milonni , it the quantum vacuum , academic press , san diego , 1994 ; m. bordag , u. mohideen , and v. m. mostepanenko , phys
. rep . * 353 * , 1 ( 2001 ) ; k. a. milton , _ the casimir effect : physical manifestations of the zero - point energy _
( world scientific , singapore , 2001 ) ; s. reynaud _
et al . _ , c. r. acad .
paris * iv-2 * , 1287 ( 2001 ) ; k. a. milton , j. phys . a : math
. gen . * 37 * , r209 ( 2004 ) ; s.k .
lamoreaux , rep .
68 * , 201 ( 2005 ) ; m. bordag , g.l .
klimchitskaya , u. mohideen , and v. m. mostepanenko , _ advances in the casimir effect _ , oxford university press , oxford , 2009 .
n. d. birrell , p. c. w. davies , `` quantum fields in curved space , '' cambridge university press ( 1982 ) ; s.a .
fulling , aspects of quantum field theory in curved spacetime " , london mathematical society student texts ( 1989 ) ; l. parker and d. toms , quantum field theory in curved spacetime : quantized fields and gravity " , cambridge university press ( 2009 ) .
the adiabatic order is usually defined as the number of derivatives of the metric tensor . in this case , and as discussed in ref.@xcite , the definition is generalized as follows : in terms containing derivatives of the classical field @xmath1 , the adiabatic order is the number of derivatives of @xmath1 plus the exponent of @xmath1 . | it is well known that boundary conditions on quantum fields produce divergences in the renormalized energy - momentum tensor near the boundaries .
although irrelevant for the computation of casimir forces between different bodies , the self - energy couples to gravity , and the divergences may , in principle , generate large gravitational effects .
we present an analysis of the problem in the context of quantum field theory in curved spaces .
our model consists of a quantum scalar field coupled to a classical field that , in a certain limit , imposes dirichlet boundary conditions on the quantum field .
we show that the model is renormalizable and that the divergences in the renormalized energy - momentum tensor disappear for sufficiently smooth interfaces . |
we consider a two - dimensional scalar field @xmath2 defined on the upper - half - plane , @xmath3 , which is free in the two - dimensional bulk but subject to a periodic boundary potential .
the action is @xmath4 where @xmath5 is a complex parameter which dials the strength and the phase of the periodic boundary interaction .
the period of the potential is chosen such that the interaction has dimension one under boundary scaling .
this ensures that the boundary theory preserves half of the conformal symmetry of the bulk system .
this model arises in various contexts , for example , in connection with critical behavior in dissipative quantum mechanics @xcite , quantum hall edge states @xcite , and open string theory in an on - shell tachyon background @xcite .
more recently a @xmath6 boundary conformal field theory involving a scalar field with a ` wrong - sign ' kinetic term and an exponential boundary interaction has been applied to the rolling tachyon field of an s - brane or an unstable d - brane @xcite .
such a theory is related by analytic continuation of the field to the boundary theory considered here .
a conformal field theory analysis of the model ( [ action ] ) , carried out a decade ago @xcite , revealed an underlying @xmath1 symmetry which allows many exact results to be established , including exact boundary states and the one - loop partition function . in this lecture
we report on ongoing work aimed at calculating correlation functions of both bulk and boundary fields .
the general problem involving fields carrying arbitrary momenta remains unsolved but we give a prescription for correlation functions of an important sub - class of fields , the so - called discrete primaries .
we take the scalar field to be non - compact , _
i.e. _ taking value in @xmath7 , although the compact case , where the field takes value on a circle of radius @xmath8 , is also of interest . in the absence of the boundary interaction , the field @xmath2 satisfies a neumann boundary condition at @xmath9 , and this is conveniently dealt with using the so - called doubling trick or method of images ( see f.ex .
the theory on the upper half - plane is then mapped into a chiral theory on the full complex plane and the boundary eliminated . to see how this works we note that away from the boundary the free field can be written as a sum of left- and right - moving components , @xmath10 . for @xmath11 the neumann boundary condition on the real axis @xmath12 determines the right - moving field in terms of the left - moving one .
we can therefore reflect the right - moving field through the boundary and represent it as a left - moving field in the lower - half - plane .
the theory then contains only left - moving fields , but a left - moving field at @xmath13 in the unphysical lower - half - plane is to be interpreted as a right - moving field at @xmath14 in the physical region . and @xmath13 denote the complex conjugate of @xmath14 .
we use @xmath15 for the argument of a right - moving field @xmath16 , but @xmath13 for that of a left - moving image field @xmath17 . ] more generally , any quasi - primary field in the bulk separates into left- and right - moving parts @xmath18 . using the doubling trick @xmath19
becomes a left - moving field @xmath20 , with holomorphic dimension @xmath21 .
a bulk @xmath22-point function @xmath23 in the original theory on the upper - half - plane then becomes a @xmath24-point function of holomorphic fields @xmath25 on the infinite plane .
it turns out that the doubling trick can be applied even when the boundary interaction in ( [ action ] ) is turned on .
this is because the boundary potential can in fact be expressed in terms of the left - moving field alone @xmath26 the operators appearing in the interaction are currents of a left - moving @xmath1 algebra @xmath27 in the boundary action ( [ action ] ) the @xmath28 and @xmath29 currents are integrated along the real axis and in a perturbative expansion of a correlation function such integrals are repeatedly inserted into the amplitude . as usual , divergences arise when operator insertions coincide but , by a clever choice of regularization , callan @xcite were able to sum the perturbation series explicitly to obtain the exact interacting boundary state .
it turned out to be remarkably simple , with the net effect of the interaction being a global @xmath1 rotation , @xmath30 , acting on the free neumann boundary state .
the bulk theory is that of a free boson .
it contains holomorphic primary fields @xmath31 with conformal weight @xmath32 for all @xmath33 and also the corresponding anti - holomorphic fields . at special values of the momentum , @xmath34 , where @xmath35 is an integer or integer - plus - half , some descendant states have vanishing norm and new primary fields appear , the so - called discrete primaries @xcite , which come in @xmath1 multiplets labelled by @xmath35 and @xmath36 , with @xmath37 .
the discrete fields @xmath38 in a given @xmath1 multiplet are degenerate in that they all have conformal weight @xmath39 .
they are composite fields made from polynomials in @xmath40 , @xmath41 , _ etc .
_ accompanied by @xmath42 , and normal ordered with respect to the free holomorphic propagator @xmath43 .
the discrete fields have the following representation @xcite , @xmath44 where the lowering current is integrated along nested contours surrounding @xmath14 .
a discrete bulk primary is constructed from a pair of holomorphic and anti - holomorphic primaries , @xmath45 a priori , the left- and right - moving fields can carry different @xmath1 labels .
however , the spin @xmath46 takes an unphysical value unless @xmath47 .
we are considering a non - compact free boson , which has no winding states , so we must also require @xmath48 , which amounts to @xmath49 .
finally we have @xmath50 where @xmath51 .
when a bulk field approaches the boundary at @xmath52 , new divergences appear that are not removed by the bulk normal ordering .
this is a general feature of conformal field theories with boundaries and signals the presence of so called boundary operators .
the boundary conditions to either side of a boundary operator can be different , in which case it is called a boundary condition changing operator ( see for example @xcite ) . in a general boundary
conformal field theory a bulk field approaching the boundary can be expanded in terms of boundary fields @xcite , @xmath53 where @xmath54 .
the @xmath55 are boundary fields ( possibly boundary condition changing ) with boundary scaling dimensions @xmath56 and the @xmath57 are called bulk - to - boundary operator product coefficients .
in addition to the bulk - to - boundary ope , the boundary fields form an operator product algebra amongst themselves , @xmath58 the boundary ope coefficients @xmath59 and the bulk - to - boundary ope coefficients @xmath57 , along with the boundary scaling dimensions @xmath56 , are characteristic data of a given boundary conformal field theory .
in particular , the boundary scaling dimension @xmath56 of a boundary operator @xmath55 is given by the energy eigenvalue of the corresponding open string state , where the open - string hamiltonian is the @xmath60 generator of the virasoro algebra that is preserved by the conformally invariant boundary conditions .
in the theory at hand , boundary conditions are labelled by the boundary coupling @xmath5 in ( [ action ] ) .
a boundary condition changing operator that changes @xmath5 to @xmath61 at the insertion point corresponds to an open string state where the two string endpoints interact with boundary potentials of different strength @xmath5 and @xmath61 . to work out the open string spectrum we find it convenient to re - express the interacting boundary theory in terms of free fermions as shown in @xcite .
the @xmath1 currents are bi - linear in the fermions and the boundary interaction may be viewed as a localized mass term .
the open string spectrum is then found by solving a straightforward eigenvalue problem for the fermions .
we will not repeat the construction here but simply quote the result . in @xcite
both string endpoints were taken to interact with the same boundary potential , _
i.e. _ it was assumed that @xmath62 . in this case
the partition function may be written @xmath63 where @xmath64 is the dedekind eta function , @xmath65 with @xmath66 the parameter length of the open string , and @xmath67 is related to the target space momentum @xmath68 by @xmath69 the value of @xmath67 is determined by continuity from @xmath70 at @xmath11 .
the scaling dimensions of the corresponding boundary operators are given by @xmath71 with @xmath72 .
the spectrum obtained from ( [ ggspectrum ] ) is shown in the left - most graph in .
the free spectrum , @xmath73 , has split into bands with forbidden gaps in energy in between them .
note that the energy eigenvalues are invariant under @xmath74 for any integer @xmath22 .
the boundary potential breaks translation invariance in the target space to a discrete subgroup and target space momentum is only conserved mod @xmath75 in our units in the interacting system .
the momentum of a given operator can therefore always be shifted into the so - called first brillouin zone , @xmath76 .
the spectra in are displayed in an extended zone scheme , with the periodicity appearing explicitly .
it is interesting to note that bands also appear in the open string channel of the boundary conformal field theory of a free boson compactified on a circle , when the radius of the circle is an irrational multiple of the self - dual radius @xmath77 @xcite . like the system we are considering here ,
those theories admit a one - parameter family of boundary states that interpolate between neumann and dirichlet boundary conditions , but it is unclear to us at present how deep the parallels between the systems run .
the band spectrum described by ( [ ggspectrum ] ) is a special case of a more general structure that appears when we allow for open strings with different boundary coupling , @xmath5 and @xmath61 , at the two endpoints . for simplicity
we take @xmath78 , with @xmath79 .
the fermion eigenvalue problem solved in @xcite can easily be extended to cover this case also .
the only modification is to change @xmath5 to @xmath61 in one of the boundary mass terms for the worldsheet fermions in equation ( 29 ) of that paper , leaving the other one unchanged .
the eigenvalue equation for the spectrum then becomes @xmath80 where @xmath81 .
clearly this reduces to the previous result ( [ ggspectrum ] ) as @xmath82 but when @xmath83 there are important new features .
in particular , there are additional gaps in the spectrum as shown in .
but the other one takes the values a ) @xmath84 , b ) @xmath85 , c ) @xmath86 .
[ fig : genspectrum ] ]
in this section we consider scattering amplitudes involving bulk fields .
general bulk amplitudes are functions of the boundary coupling @xmath5 and our goal is to determine this dependence .
some bulk amplitudes that are zero in the free theory are nonvanishing in the presence of the boundary interaction . the periodic boundary potential breaks translation invariance in the target space and can absorb momenta in @xmath75 .
callan @xcite gave a simple prescription for scattering amplitudes that describe elementary string excitations reflecting off the interacting worldsheet boundary .
their method , which can be called _ the method of rotated images _
, rests on the fact that the bulk operators @xmath87 and @xmath88 , which create and destroy left- and right - moving excitations , are in fact currents of the underlying @xmath1 algebra .
the method can be generalized to deal with bulk scattering amplitudes involving operators , which carry well defined @xmath1 quantum numbers , _
i.e. _ the discrete bulk operators @xmath89 in equation ( [ discretebulkfield ] ) . consider some number of these states inserted in the upper half - plane and use the method of images so that for each insertion @xmath90 .
this is possible even with the boundary interaction turned on because the anti - holomorphic field @xmath91 commutes with the holomorphic @xmath1 currents in the interaction and reflects through the underlying neumann boundary condition exactly as in the free theory . in a perturbative expansion of the bulk amplitude the @xmath1 currents that appear in the boundary interaction
are repeatedly integrated along the real axis .
their integration contours may be deformed away from the real axis into the lower - half - plane , where they will act on the image fields @xmath92 .
the @xmath1 current algebra ensures that there will be no cuts generated , and so the integration contours can be closed around the image fields .
the net effect is the global @xmath1 rotation @xmath93 , given at the end of section [ currentalgebra ] , acting on each image field insertion .
the original right - moving discrete field was a component of a rank @xmath94 irreducible tensor operator , so the rotated image is @xmath95 with the rotation coefficient given by @xmath96 where @xmath97 are standard @xmath1 states . a bulk @xmath22-point function
is then expressed in terms of @xmath24 point functions of free holomorphic fields and @xmath1 rotation coefficients , @xmath98 as a simple example of this prescription we consider the one - point function of a general discrete bulk primary , @xmath99 conformal invariance requires the scaling dimension of the two chiral operators to be the same , i.e. @xmath100 , and momentum conservation in the free chiral theory requires @xmath101 .
the one - point function is therefore @xmath102 for comparison , note that in the free theory with neumann boundary conditions the only bulk operator that has a non - vanishing one - point function is the unit operator .
higher - point bulk amplitudes involving discrete fields can be computed in an analogous fashion , but unfortunately our prescription can not be applied to all bulk amplitudes in the model . in general , bulk amplitudes will involve both the discrete bulk fields and fields carrying generic target space momenta .
the only constraint from momentum conservation is that the total momentum of all the fields in a given correlator add up to an integer times @xmath103 in our units .
the operator product of the currents in the boundary interaction and generic momentum fields is non - local , and so the effect of the interaction is no longer captured by a global @xmath1 rotation .
the physics of the open string sector is encoded in correlation functions of boundary operators as discussed in section [ boundaryfields ] .
it is straightforward to identify the boundary operators in the free theory .
consider bringing a bulk primary operator at generic momentum @xmath104)$ ] close to the boundary . now
replace the right - moving part by its left - moving image and apply the left - moving ope , @xmath105 , where @xmath106 and the boundary scalar field is related to the left - moving field by @xmath107 .
for the discrete bulk fields one obtains @xmath108 the @xmath109 are primary fields on the boundary , which we refer to as _ discrete boundary fields_. the superscript on @xmath110 signals that these are boundary operators of the free theory . the bulk - to - boundary
ope coefficients @xmath111 can be obtained by straightforward calculation . by momentum conservation
they vanish unless @xmath112 .
the discrete boundary fields inherit the @xmath1 structure from the chiral discrete fields , but since @xmath113 we find that only integer values of @xmath114 are allowed on the boundary .
the boundary scaling dimension of @xmath110 is @xmath115 .
we are interested in calculating amplitudes involving arbitrary boundary fields in the interacting theory , including boundary condition changing fields .
this is delicate since operators are now inserted on the boundary where the non - linear self - interaction takes place .
we can nevertheless anticipate the structure of low - order amplitudes from conformal symmetry .
for two- and three - point functions one finds @xmath116 is the scaling dimension obtained from equation .
it remains to determine the @xmath5-dependence of the coefficients @xmath117 and @xmath59 .
explicit calculations for fields carrying arbitrary momenta are difficult because such fields are non - local with respect to the @xmath1 currents in the boundary interaction , but we can proceed further with amplitudes involving the discrete boundary fields .
a key observation , that can be read off from equation , is that the boundary scaling dimension of @xmath118 is @xmath115 , independent of the coupling . as a result
we can write @xmath118 as a linear - combination of free boundary operators within the same @xmath1 multiplet @xmath119 we now write the general bulk - to - boundary ope at non - zero boundary coupling , @xmath120 we then apply the method of rotated images as described in section [ bulkamplitudes ] on the left - hand side , plug in the expansion ( [ eq : psig ] ) on the right - hand side , and use equation ( [ eq : dbb ] ) .
this results in a set of algebraic equations that relate the ope coefficients @xmath121 and the expansion coefficients @xmath122 .
unfortunately , there are not enough equations to determine all the coefficients , so we need further input .
we propose the following prescription for determining @xmath122 .
first the free boundary operators @xmath110 are obtained as in equation .
the effect of the interaction on these operators is then computed by letting the integration contours of the boundary currents approach the real axis , where @xmath110 is inserted , from above .
the contours are then moved into the lower - half - plane , resulting in an @xmath1 rotation acting on @xmath110 , giving @xmath123 .
this leads to boundary amplitudes of the form @xmath124 and bulk - to - boundary ope coefficients @xmath125 we stress that other prescriptions are possible .
one alternative would be to deform the integration contours into the upper - half - plane in which case @xmath126 and the effect of the interaction is shifted entirely into the bulk - to - boundary coefficients @xmath121 .
we note , however , that the boundary amplitudes ( [ boundaryamplitudes ] ) have the desirable feature that momentum is only conserved modulo @xmath103 which reflects broken translation invariance .
we thank c. callan , d. friedan and i. klebanov for discussions .
this work was supported in part by grants from the icelandic research council , the university of iceland research fund and the icelandic research fund for graduate students .
thanks the new high energy theory center at rutgers university for hospitality . | correlation functions of discrete primary fields in the @xmath0 boundary conformal field theory of a scalar field in a critical periodic boundary potential are computed using the underlying @xmath1 symmetry of the model .
bulk amplitudes are unambigously determined and we give a prescription for amplitudes involving discrete boundary fields . |
several recent efforts have been done in order to understand the subyacent dynamics in electoral systems and opinion formation @xcite , from the contrarian effect@xcite to the , so called , small world behavior@xcite . in the brazilian elections , for instance , power law was found for the proportional vote@xcite .
however , to establish a proper description of an electoral process is a hard issue since many factors and interactions appear and several aspects of them must be studied
. statistical characterization of actual processes is an important issue as well , mainly with the increasing possibility of obtaining the vote data . in the present work
we incorporate the analysis on the corporate vote with a study on the statistical properties of the federal mexican elections of 2000(e-2000 ) , 2003(e-2003 ) , and 2006(e-2006 ) . since their distributions are smooth , the existence of an analytical distribution that describes them and a model which explains them , are very tempting issues .
we were successful in the first topic but the answers to the second remain open .
we find a remarkable well fit of the properly unfolded distribution of votes with a family of distributions obtained in the context of spectral statistics of complex quantum systems , the called daisy models @xcite .
the process presented here is different from those that appears in @xcite , for instance , since the vote decision is taken due to pertain to a corporate . in the referred electoral processes
two new features appeared : i ) the party who ruled for around @xmath2 years@xcite became opposition and ii ) the vote data are public and in an electronic format@xcite .
the last fact allows an extense statistical analysis , meanwhile , the former one gave the opportunity to analyze the vote distribution of the hard core or corporate voters .
we shall denote this party as p2 according to the place in which it appears in the data basis@xcite .
here we present the vote distribution of this party during the elections 2000 , 2003 , and 2006 for the president position and places in both chambers .
this distribution consist of the histogram of the number of votes per cabin obtained for each party , i.e. , in how many cabins exists @xmath3 vote , @xmath1 votes and so on .
the crude data histograms are presented in the next section , and the unfolding procedure together with the theoretical distribution appear in section [ unfolding ] followed by some remarks and conclusions .
in a recent analysis of the mexican election of 2006@xcite the distributions of votes for all the participants were done with data obtained from the preliminary results program ( prep ) .
for the two main parties the results are unclear and a mix of processes is expected ; meanwhile for the rest of the forces clear distributions appear : power laws for small parties , annulled votes and non - registered candidates , and a smooth distribution for the third electoral force .
the corporate party p2 present a histogram with a clear maximum and a tail with exponential decay . in fig .
[ fig:1](a ) the histograms of the official final results for p2 are presented .
the data were obtained from the electoral authorities web page @xcite and on request @xcite . by construction ,
each cabin admits only @xmath4 votes from the registered list of voters and they are distributed over all the country being a sample grid on the population aged over @xmath5 years old@xcite . as an exception
, there exist special cabins for voters in transit but the number of them is small and do not affect the statistical results presented here
. with those remarks , it is clear that the histograms are statistics .
the analysis of the link with the geo - economic regions is beyond of the present work but it is of interest . in fig .
[ fig:1](a ) and [ fig:1](c ) the histograms for presidential ( red ) , deputies ( black ) and senators ( blue ) for e-2000 ( upper panel ) and e-2006 ( lower panel ) are presented and the intermediate elections for the low chamber in e-2003 ( fig .
[ fig:1](c ) in violet ) .
we present the crude data with no average of any sort .
other parameters of the distributions are presented in table [ table:1 ] being the total number of cabins considered in each process the following : @xmath6 for e-2000 , @xmath7 for e-2003 and @xmath8 for e-2006 .
the existence of a smooth distribution that fits the data is a very tempting issue and is the matter of the rest of the present work . .
[ cols="<,>,>,>",options="header " , ]
a direct comparison with any probabilistic distribution requires of proper normalization and unfolding of the signal , i.e. separate the secular part from the fluctuating one . to consider this procedure is important since in many cases considering the relative variable @xmath9 is not enough because the average @xmath10 could not be constant through the whole set of data . in general , the unfolding procedure is a very delicate task @xcite . in the present case no _ a priori _ density can be defined , since it is not clear if the alphabetical order in which the data basis is ordered corresponds to the dynamics of the system . to test the fitting ,
two sorts were considered , the original order and a randomly sorted sequence of the votes . in both cases
the histograms are similar but the last one gives much more stable results after the unfolding procedure , as we shall discuss below . in order to fit the experimental data with a probability distribution we treat the number of votes @xmath11 as if they were differences of energies in a quantum spectrum . as in the case of energy levels
, we consider the spectrum,@xmath12 , formed by levels @xmath13 where @xmath11 is the number of votes in the cabin @xmath14 and we define @xmath15 .
it is costumary to consider the integrated spectral function or integrated density @xmath16 , which counts the number of levels @xmath17 with value equal or less that @xmath18 .
@xmath19 stands for the heaveside unitary step function .
the integrated density is decomposed into a secular @xmath20 and a fluctuating part @xmath21 .
the former part is given by the integral of the correlation function of one point ( see ref .
@xcite for explanation ) .
the sequence @xmath17 is mapped onto the numbers @xmath22 , with @xmath23 .
the new variable is the unfolded one which has a constant density and the statistical analysis is performed on it . in the case of quantum systems
@xmath20 is estimated applying semiclassical rules .
the first term of its expansion is called , in the literature , the thomas - fermi estimate or , the weyl term in the case of billiards . in practical situations
this function is evaluated via polynomial fitting .
the last procedure described was done in the present case since it is the standard in many fields of physics and it is of general applicability .
the specific statistic that we analyze corresponds to the nearest neighbor spacing , @xmath24 , defined as @xmath25 and is the unfolded version of the number of votes @xmath11 defined in eq .
( [ intvote ] ) .
the results for the unfolded votes @xmath26 are shown in figs .
[ fig:2 ] and [ fig:3 ] for the deputies , president and senators in 2000 and 2006 elections . for sake of clarity
we drop the e-2003 analysis .
the theoretical comparison is made with the daisy model of rank @xmath27 @xcite .
this model departs from retaining each @xmath0 level from a set of levels with a poisson distribution .
the resulting sequence has the @xmath0-nearest neighbor distribution of poisson s , but it must be renormalized in order to obtain the nearest neighbor distribution of the daisy model of rank @xmath27 .
the resulting @xmath28-th neighbor spacing distribution is @xmath29n)}s^{(r+1)n-1}\exp[-(r+1)s ] .
\label{rdaisy}\ ] ] where @xmath27 corresponds to the kind or rank of the family .
the rank @xmath30 corresponds to the semi - poisson distribution which is related to the energy distribution of pseudointegrable systems@xcite and others systems at criticality@xcite like as the disordered conductor at the anderson transition @xcite . a strong relation exists between daisy models and the nearest - neighbor interaction one - dimensional coulomb gas@xcite , where the dependence in the inverse temperature @xmath31 from the later model has the same role as the rank @xmath27 in the former . we do no have an _ a priori _ density in order to compare the vote records with the distribution of eq .
( [ rdaisy ] ) .
then , the present study can be done only at nearest neighbors , @xmath32 , even when an exploration to larger range correlations is extremely interesting and will be matter of future works .
the model described by eq .
( [ rdaisy ] ) with @xmath32 fits the unfolded p2 distribution of vote in two regions with two different daisy ranks , @xmath27 , the central part is usually fitted by a higher rank and decays with a lower one . in the case of e-2000 ( fig .
[ fig:2 ] ) the fit is between @xmath33 and @xmath34 , but a remarkable deviation is that the experimental distributions start with a linear grow as indicated below and with the dashed line in fig . [ fig:1](a ) .
this characteristic remains after the unfolding procedure and marks a clear deviation from the @xmath35 behavior , however , if we do not consider this data , the area preservation of the distribution makes that the function with @xmath34 fits better .
the decay is well fitted by a @xmath36 daisy model in all the cases , as can be seen in the lower panel of fig .
[ fig:2 ] .
it is important to note that a fit to weibull / brody distribution is not good , since the decay is clearly exponential and that only happens in the poissonian case of such distribution .
clearly this is not the case . in e-2006
a differential vote occurs since the presidential candidate obtains around @xmath37 less votes than p2 obtains for the chambers , as can see in table [ table:1 ] .
such an event does not happens in the e-2000 process . in this case
, the vote distributions for the chambers fit with a @xmath38 for the whole range , the body and the tail ( fig . [ fig:3 ] ) . for the presidential case
exists a general fit with the @xmath38 for the body , but decays with a rank @xmath1 .
note that the agreement for the presidential case ( red histogram ) is not as good as the other cases .
the main difference in e-2000 and e-2006 is that p2 arrived to the last process with a deep internal division as was widely reported in national newspapers .
an interesting remark , since the votes distribution has fit different rank of daisy models is that the parameter @xmath0 plays the role of an inverse temperature when the these models are contrasted with the statistical distributions of a @xmath3-dimensional coulomb gas with logaritmic interactions @xcite
in this work we presented an analysis of the vote distribution for one of the corporate parties ( named p2 here ) in mexico which has a wide influence in all the country since it was in the federal power for around @xmath2 years . by construction
the mexican electoral system admits a straight statistical analysis since the number of cabins are defined and distributed in order that each cabin admits only @xmath4 voters .
the crude p2 vote distributions look smooth ( fig .
[ fig:1 ] ) and , after a proper unfolding and normalization , corresponds to a probability distribution .
comparison of the data with nearest neighbor daisy models of rank @xmath27 ( @xmath32 ) gives a good agreement in all the cases . in the election in 2000
the agreement could be better , but the data distribution depart from a linear grow as indicated in fig .
[ fig:1](a ) with a dashed line tor guide the eye .
the distributions tails follow a different daisy model rank .
the dynamical meaning of these results is unclear but the agreement with a daisy model suggest the existence of universal processes therein and not just a fortuitous agreement . in ref .
@xcite the daisy model of rank @xmath1 fit the distribution of distances for the quasi - optimal path in the traveling salesman problem ( tsp ) .
the problem consists in finding the shortest path between @xmath39 cities visiting each city just one time .
it is clear that the problem presented here is of the same type .
p2 have voters in each region of the country and their conform a truly wide web .
how this happens is matter of future analysis , as well the existence of similar behavior in other corporate parties around the world .
this work was partially supported by dgapa - unam project in-104400 and promep 2115/35621 .
oo a. carbone , g. kaniadakis , and a. m. scarfone .
j. b. * 57 * 121(2007 ) .
this reference offers a large but non - exhaustive list of references .
s. fortunato and c. castellano , phys .
lett . * 99 * , 138701 ( 2007 ) . c. borghesi and s. galam , phys .
e * 73 * , 066118 ( 2006 ) . a large amount of existing references can be obtained in science citation index .
gonzalez , a.o .
sousa and h.j .
c * 15 * , 45 ( 2004 ) .
costa filho , m.p .
almeida , phys .
e , * 60 * , 1067 ( 1999 ) ; physica a * 322 * , 698 ( 2003 ) .
a.a . moreira , d.r .
paula , r.n .
costa filho and j.s .
andrade jr .
rev e. * 73 * , 065101(r ) ( 2006 ) . h. hernndez - saldaa , j. flores and t.h .
e. * 60 * , 449 ( 1999 ) . ` http://www.ife.org.mx ` according with the mexican transparency law , the electoral data are public and can be obtained on request .
the party p2 is the partido revolucionario insitutional ( pri ) .
g. baz , h. hernndez - saldaa and r.a mndez - snchez . to be submitted to phys .
e. e - print ` arxiv : physics/0609114 ` to be registered in the voters list is almost mandatory in mexico since is required as the official i d card for all practical purposes .
t. guhr , a .
mller - groeling , and h.a .
phys . rep . *
299 * , 190 ( 1998 ) .
g. casati and t. prosen .
lett . * 85 * , 4261 ( 2000 ) .
bogomolny , u. gerland , and c. smith .
e. * 59 * , r1315 ( 1999 ) ; eur .
j. b. * 19 * , 121 ( 2001 ) .
u. gerland , ph .
d. thesis , university of hilderberg , germany , 1998 .
a.m. garca - garca and j. wang .
e. * 73 * , 036210 ( 2006 ) .
shklovskii , b. shapiro , b.r .
sears , p. lambrianides , and h.b .
shore , phys . rev . * 47 * , 11487 ( 1993 ) . | the distribution of votes of one of the corporate parties in mexico during elections of 2000 , 2003 and 2006 is analyzed .
after proper normalization and unfolding , the agreement of the votes distributions with those of daisy models of several ranks is good .
these models are generated by retaining each @xmath0 level in a sequence which follows a poisson distribution .
beyond the fact that rank @xmath1 daisy model resembles the distribution of the quasi - optimal distances for the traveling salesman problem , no clear explanation exists for this behavior , but the agreement is not fortuitous and the possibility of a universal phenomena for corporate vote is discussed . |
the radio pulsar psr b1259 @xmath063 is in a highly elliptical orbit about the luminous be star ss2883 .
pulsar winds are expected to accelerate electrons to lorentz factors of up to @xmath1 , leading to up - scattering of the ultra - violet photons from the be star into the tev range @xcite .
this can happen before or after the wind passes through its termination shock @xcite .
however , in the case of psr b1259 @xmath063 , the time spent by an individual electron in the unshocked wind is short compared to the time spent in the vicinity of the be star after passing the shock .
thus , even if the shock simply isotropises the electrons without energising them , the post - shock emission should dominate over the pre - shock emission .
this conclusion holds a forteriori , if , as expected , the shock transfers some of the incoming kinetic energy into nonthermal particles .
observations using the h.e.s.s .
array of imaging erenkov telescopes around and after the periastron passage in early 2004 detected a strong signal in the tev range @xcite .
the measured spectrum is in excellent agreement in both slope and absolute normalisation with that predicted by a model in which the post - shock pulsar wind electrons have a simple , single power - law distribution @xcite .
significant night - to - night fluctuations in the tev light curve as well as an overall decrease on the timescale of months were also observed by h.e.s.s .
, possibly correlated with variations in the unpulsed radio emission @xcite .
short timescale fluctuations , especially close to periastron , can plausibly be attributed to departures from spherical symmetry in the structure of the pulsar wind or the be star wind ; the most detailed current model ascribes the variation in the unpulsed radio emission to the latter @xcite .
these were not taken into account in the predictions of the tev emission @xcite , which included only those effects arising from the variation of the stellar separation over the orbit . a model in which the tev emission arises from proton - proton interactions in the anisotropic wind ( disk ) of the be star @xcite produces short timescale features qualitatively similar to those observed , but appears to predict a flatter tev spectrum than that seen by h.e.s.s . in this paper , we present preliminary results from an extended version of the model of @xcite .
injection of a double power - law electron spectrum , similar to that thought to be injected into the crab nebula by its central pulsar @xcite is included , as is the transition from radiative to adiabatic loss mechanisms as the stellar separation increases .
it has recently become clear @xcite that the pulsar wind that fuels the crab nebula injects into it relativistic electrons and positrons ( and possibly ions ) whose energy distribution can be approximated as a double power - law : @xmath2 , for @xmath3 and @xmath4 for @xmath5 .
the high - energy index is determined by the slope of the x - ray spectrum of the crab nebula : @xmath6 , in agreement with theories of first - order fermi acceleration at relativistic shocks @xcite .
the low energy index follows from the slope of the radio to infra - red spectrum : @xmath7 . with these values
, most particles are concentrated around the lower cut - off at @xmath8 , whereas most of the energy is injected in electrons of lorentz factor @xmath9 . in the crab , @xmath10 , @xmath11 and @xmath12 .
the resulting synchrotron spectrum contains two breaks , one due to cooling and one intrinsic to the injected spectrum ( at @xmath13hz and @xmath14hz in the crab ) , as well as upper and lower cut - offs . if this injection spectrum is produced at the termination shock front , and if the cold upstream flow is dominated by the kinetic energy flux in electron - positron pairs , then the lorentz factor of the wind is @xmath15 . in the following we adopt this injection model . in psr b1259 @xmath063
, relativistic electrons and positrons in the shocked pulsar wind suffer adiabatic losses as the plasma expands away from the shock front , as well as radiative losses by synchrotron radiation and inverse compton scatterings , primarily of the ultra - violet photons from the be star .
the energy dependence of these loss processes is different and influences the resulting distribution function .
two sets of models were constructed in @xcite : one for dominant adiabatic and one for dominant radiative losses .
both were calibrated using the observed x - ray synchrotron emission , and provided accurate predictions of the tev spectrum subsequently detected just before periastron .
however , the two models imply quite different injection spectra . as the pulsar moves away from the be star , both the target radiation field and the magnetic field where the winds interact decrease , along with the gas pressure .
for the toroidal field structure expected in a pulsar wind , the ratio of the energy densities of magnetic field and target radiation remain constant , so that , in the absence of klein - nishina effects , the ratio of synchrotron to inverse compton radiation should not vary with binary phase .
however , if the expansion time scales linearly with the stellar separation , adiabatic losses become more important with respect to radiative losses as the stars move apart . in the models used here ,
we account for this by switching between a radiative and an adiabatic loss term in the kinetic equation of the electrons at the lorentz factor where the loss rates coincide .
the losses themselves are specified by the magnetic field strength in the emission region , and the adiabatic loss time scale at periastron .
l|lllllll model:&@xmath16 & @xmath17&@xmath18&@xmath19&@xmath20&efficiency&@xmath21 + & & a & @xmath22 & @xmath1 & @xmath23 & @xmath24 & @xmath25 g & @xmath26% & 15 + b & @xmath22 & @xmath27 & @xmath28 & @xmath29 & @xmath25 g & @xmath30% & 0.5 + the emitted radiation is a combination of synchrotron radiation in a uniform magnetic field and inverse compton scattering of ultra - violet photons from the be star . on its way from the pulsar system to the observer
the inverse - compton emission is partially reabsorbed via pair production on the stellar photons @xcite .
these processes are well - understood . to expedite the computations , @xcite used delta - function approximations to the emissivities of both emission mechanisms . in addition to the standard assumption ( of synchrotron theory ) that the direction of the emitted photons is approximately that of the incoming electron
, these approximations replace the energy spread of photons emitted by a given electron by an appropriate monochromatic term . on physical grounds ,
this approximation can be expected to give better results as the electron lorentz factor increases and to be especially accurate in the klein - nishina regime of inverse compton scattering .
we have checked these approximations using a computationally more costly evaluation of the full klein - nishina rates and find they indeed give very accurate results , in conflict with the comparison presented in @xcite .
on the other hand , care must be taken with these approximations for synchrotron radiation , especially when sharp gradients in the distribution function are present .
this is because the synchrotron process is equivalent to a scattering deep inside the thomson regime , leading to a relatively broad emission cone . in this paper
we use the full synchrotron emissivity , but keep the delta function approximation for inverse compton scattering used in @xcite .
modelling the spectrum and light curve of the high energy emission during the 2004 periastron passage is made difficult by the scarcity of simultaneous tev and x - ray data sets .
the only ones available are the x - ray / soft gamma - ray spectrum detected by integral between 14 and 17 days after periastron passage @xcite and the march 2004 observations by h.e.s.s .
@xcite . these data are not sufficient to determine the dominant loss mechanism . to illustrate this , we consider two models , with parameters given in table [ parameters ] , where we also quote the efficiency of each model , defined as the ratio of the power injected into the emission region in the form of relativistic electrons to the pulsar spin - down power .
the tev gamma - ray spectrum implies that the differential number of radiating electrons is roughly @xmath31 .
this can be modelled either as the result of radiative cooling of the hard injection spectrum at @xmath32 , or of the accumulation without energy loss of electrons injected at @xmath33 .
the first case applies if radiative losses dominate .
the second , if the losses are adiabatic .
spectra and light - curves for model a ( radiative - loss dominated ; left panels ) and model b ( adiabatic - loss dominated ; right panels ) together with observations by integral ( green ) @xcite and h.e.s.s .
( blue ) @xcite are shown in fig .
[ spectrum ] .
the vertical blue lines depict the photon energies associated with electron lorentz factors at which the adiabatic and radiative cooling rates are equal . at these points ,
a cooling break appears in the spectrum .
the vertical green lines mark the break intrinsic to the injection function , at @xmath34 .
the dotted curve indicates the intrinsic emission , before propagation through the radiation field of the be star .
it is evident that model a , can not produce as good a fit to the ( hard ) x - ray spectrum as model b , in which adiabatic losses dominate .
the reason is that , in the hard x - ray range , cooling by synchrotron emission is more important than cooling by inverse compton emission . in this case
, the hardest possible model spectrum produced by cooled electrons has a photon index of @xmath35 .
this limitation does not apply to the adiabatic - loss dominated models .
however , the data are not sufficient to reject model a , because of the relatively large error associated with the spectrum reported by integral ( @xmath36 ) . it should be noted that in _ soft _ x - rays , radiative - loss dominated models produce a harder spectrum .
this interesting effect , ( remarked upon by many authors including @xcite ) arises from the transition between inverse compton cooling at low frequencies and synchrotron cooling at higher frequencies as the scattering regime moves from thomson into klein - nishina . the key property of the models that the tev spectrum is formed by electrons of @xmath32 in model a ( radiation losses ) and @xmath33 in model b ( adiabatic losses ) implies a faster pulsar wind in model b , but also a very high efficiency .
since the computed efficiency depends on the poorly known distance to this object , values in excess of @xmath37 are not forbidden . nevertheless , they are uncomfortable , since the spin - down power must also provide for the kinetic energy of the bulk post - shock flow and the poynting flux , as well as , perhaps , relativistic ions . in model
b , this problem is exacerbated by the extremely short ( @xmath38secs ) adiabatic loss time , which implies a high post - shock bulk speed . | the high - energy gamma - ray emission discovered using the h.e.s.s .
telescopes from the binary system psr b1259 @xmath063 , is modelled using an extension of the approach that successfully predicted it .
we find that the simultaneous integral and h.e.s.s .
data permit both a model with dominant radiative losses , high pulsar wind lorentz factor and modest efficiency as well as one with dominant adiabatic losses , a slower wind and higher efficiency .
additional , simultaneous , x - ray and tev data sets are needed to lift this degeneracy . |
this paper is about achieving the maximal information transfer rate possible when information is encoded into quantum systems via the preparation of one or another of two nonorthogonal states .
this might at first seem like a questionable thing to consider : for transmissions through a noiseless medium , the maximal transfer rate ( or capacity ) of 1 bit / transmission is clearly achieved _ only _ with orthogonal alphabets .
this is because nonorthogonal preparations can not be identified with complete reliability .
however , there are instances in which it is neither practical nor desirable to use such an alphabet .
the most obvious example is when a simple laser transmitter is located a great distance from the receiver .
the receiver s field will take on the character of a very attenuated optical coherent state . because the states become less orthogonal as the power is attenuated , one is confronted with precisely the issue considered here . in this case
, one is typically stuck with trying to extract information from quantum states that are not only nonorthogonal , but almost completely overlapping .
the practical method in many situations for compensating for very weak signals is to invest in elaborate receiving stations .
for instance , in microwave communication very large - dish antennas are the obvious route .
recently , however , a new quantum mechanical effect has been discovered for the decoding of nonorthogonal signals on separate quantum systems .
traditional signal processing methods have only considered fixed decoding measurements performed on the separate transmissions ( see for example ref .
@xcite ) : i.e. , taking into account the intrinsic noise generated by the quantum measurement @xcite , one is left with a basic problem of classical information theory coding for a discrete memoryless channel @xcite .
quantum mechanics , however , allows for more possibilities than this @xcite .
if one is capable of doing collective measurements on blocks of transmitted signals , it is possible to achieve a greater capacity than one might have otherwise thought @xcite this is referred to as the superadditivity of quantum channel capacities .
this is an effect that does not exist classically ( * ? ? ?
* lemma 8.9.2 ) .
the physics behind the effect relies on a kind of nonlocality dual to the famous one exhibited by entangled quantum systems through bell inequality violations @xcite .
more precisely , a communication rate @xmath0 is said to be achievable if in @xmath1 transmissions there is a way of writing @xmath2 messages with the nonorthogonal alphabet so that the probability of a decoding error goes to zero as @xmath3 .
the number @xmath0 signifies the number of bits per transmission that can be conveyed reliably from the transmitter to the receiver in the asymptotic limit .
clearly the rates that can be achieved will depend on the class of codings used for the messages and the class of quantum measurements allowed at the receiver .
the capacity @xmath4 is defined to be the supremum of all achievable rates , where @xmath5 is the number of transmissions to be saved up before performing a measurement .
the meaning of superadditivity is simply that @xmath6 , where the inequality is strict .
generally it is a difficult problem to calculate @xmath7 even with a quantum version of shannon s noisy channel coding theorem available @xcite . and
it is a much more difficult task still to find codes that approach @xmath7 .
this is because the coding theorems generally give no information on how to construct codes that approach a given capacity .
it turns out however that the number @xmath8 is rather easily calculable and , because of a recent very powerful theorem on quantum channel capacities @xcite , so is the asymptotic case @xmath9 @xcite .
the most striking thing about these two quantities is that even though both @xmath10 and @xmath11 as the overlap between the states goes to unity , the ratio @xmath12 nevertheless diverges .
this means that grossly collective measurements can , in principle at least , produce an arbitrarily large improvement in the channel capacity of very weak signals a very desirable state of affairs and one of some serious practical import .
the problem from the practical side of the matter is that before one will be able to decode very large blocks , one must first be able to tackle the case of small blocks , preferably of just size two or three
. there has already been substantial progress in this direction by sasaki _
et al_. , in a series of papers @xcite .
they explicitly demonstrate a code that uses collective decoding three transmissions at a time to achieve a communication rate @xmath13 greater than @xmath8 .
nevertheless , it would be nice to demonstrate superadditivity with an even simpler scheme , namely two - shot collective measurements . also the ratio @xmath14 as the angle @xmath15 between the two states goes to zero for their given coding scheme .
thus just where one would be looking for the most help from superadditivity ( in the very weak signal regime ) , one loses it for this code .
we improve on the work of sasaki _
et al_. , by showing that in fact @xmath16 for angles @xmath17 , and moreover that this superadditivity is sustained and only strengthened as @xmath18 . on the down side ,
the improvement in capacity is not great only 2.82 percent but is definitely there and not so small as to be forever invisible . in this vein
, we propose an experimental demonstration that relies on near - term laboratory capabilities for implementation . for our two nonorthogonal quantum states ,
we use low photon - number coherent states @xmath19 and @xmath20 with the separate signals carried on different circular polarizations .
the two - shot signal decoding is performed with atomic state measurements on a single cesium atom in a high - finesse optical cavity via the technique of quantum jumps in fluorescence similar to those demonstrated on ions in refs .
@xcite .
the remainder of the paper is organized as follows . in the next section , we demonstrate explicitly that @xmath21 for an alphabet of two nonorthogonal states and compare that to the rate @xmath13 found in refs . @xcite . in section iii
, we specialize this calculation to the coherent states mentioned above and approximate the decoding scheme of section ii with a related scheme that is first order in the small parameter @xmath22 .
finally , in section iv , we delineate the details of our experimental proposal .
following the discussion in the introduction , we will take as an alphabet for all communication schemes a fixed set of two nonorthogonal quantum states @xmath23 and @xmath24 characterized by the single parameter @xmath15 : @xmath25 we would like to know what communication rates @xmath26 can be achieved with this alphabet when decoding measurements are performed @xmath5 transmissions at a time .
this in general is a very difficult problem , especially if one is also confronted with the issue of explicitly demonstrating codes for achieving those rates .
however , if one can be contented in knowing the number @xmath7 itself and the quantum measurements required to achieve that ( i.e. , without knowing the coding scheme explicitly ) , then a great simplification arises because of a quantum extension to the shannon noisy coding theorem @xcite due to holevo @xcite .
we shall state the result of this theorem presently .
let the variable @xmath27 denote the binary strings of length @xmath5 that index the set of all messages @xmath28 denote the resultant density operator of that distribution .
we shall use the notation @xmath29 to denote a generalized quantum measurement or positive operator - valued measure ( povm ) @xcite on the message hilbert space @xmath30 , i.e. , @xmath31 is an infinite sequence of operators on @xmath30 with only a finite number of @xmath32 such that @xmath33 for all @xmath1 and @xmath34 , and the @xmath35 s form a decomposition of the identity operator on @xmath30 . in order to find @xmath7 ,
it is enough to perform the following maximization : @xmath36\ ; , \label{candy - o}\ ] ] where @xmath37 and @xmath38 are the shannon informations for the various probability distributions generated by the measurement @xmath29 .
( in these expressions we have used the base - two logarithm so that information is measured in bits . ) for any rate @xmath39 , @xmath40 , there _ exists _ a code that will achieve that rate . moreover , if @xmath29 is fixed and only the maximization over @xmath41 is performed in eq .
( [ candy - o ] ) , then the resulting expression will define the capacity that can be reached with the given measurement .
there are two limiting cases where the calculation of @xmath7 becomes tractable , @xmath42 and @xmath43 . in the first case
, one can use refs .
@xcite to find rather easily that @xmath44\log \left[1+\sqrt{1-\cos^{2}\gamma}\,\right ] \nonumber\\ & & + \frac{1}{2}\left[1-\sqrt{1-\cos^{2}\gamma}\,\right]\log \left[1-\sqrt{1-\cos^{2}\gamma}\,\right]\;. \label{littleboy}\end{aligned}\ ] ] for the limit where arbitrarily many collective measurements are made , one can use the powerful theorem of ref .
@xcite to find that the channel capacity per bit is given by @xcite @xmath45 for all cases in between , there is nothing better to be done than an explicit search over all probabilities @xmath41 and all measurements @xmath29 . as stated in the introduction
, one can see from eqs .
( [ littleboy ] ) and ( [ fatman ] ) , that @xmath46 so the incentive to use collective measurements in the decoding of these signals is great .
therefore , let us specialize to the case of collective measurements on two transmissions at a time . in this case , with respect to the decoding observables we have an effective alphabet consisting of the tensor - product states @xmath47 and
@xmath48 it turns out that these states can already exhibit superadditivity even when the collective observables are taken to be simple von neumann measurements : i.e. , by taking @xmath49 where the @xmath50 are four orthonormal vectors .
taking @xmath51 to be a probability distribution on the effective alphabet states , we must attempt to maximize the rate [ fourd ] @xmath52 with @xmath53 @xmath54 and @xmath55 and likewise for @xmath56 , @xmath57 , and @xmath58 .
the rate @xmath59 we will be interested in is then @xmath60 we have thoroughly studied @xmath61 numerically with a steepest descent and simulated annealing technique . as one might guess , the optimal solution to eq .
( [ flubber])for sufficiently small angles ( @xmath17)appears to obey the following symmetries @xmath62 therefore we make the following _ ansatz _ ( see fig .
2 ) @xmath63 [ fourc ] [ fourb ] taking these symmetries as a more analytic starting point , we can expand the measurement basis as a function of @xmath64 , @xmath15 , and the alphabet states ( see fig .
2 ) : @xmath65 thus the rate can now be expressed as @xmath66 [ foura ] even with these strong assumptions and simplifications , @xmath67 does not yield a simple analytic expression .
we must instead content ourselves with a numerical study as depicted in figure 3 .
note in particular that as @xmath68 the superadditivity does not dwindle away : @xmath69 this contrasts with the rate @xmath13 exhibited by sasaki _
et al_. @xcite for which the ratio @xmath70 goes to one within the very weak signal regime .
note that we use the notation @xmath59 rather than @xmath71 because our favored quantity can only be asserted as a _ lower bound _ to the two - shot capacity .
the symmetry assumptions on the probabilities along with the specialization to symmetric von neumann measurements could turn out to be overly restrictive .
however , further numerical investigations seem to indicate that any further improvement is likely to be very small @xcite . also we should emphasize that demonstrating that @xmath72 does not give an automatic means for finding a code that comes within @xmath73 of this rate : the channel capacity theorem eq .
( [ candy - o ] ) is only an existence proof of such a code .
however , the noise model that our alphabet and measurement leads to i.e .
, a simple stochastic transition diagram on three letters has been extensively studied in classical information - theory literature , and good codes for this problem are likely to exist .
finally , let us mention one more quantification of the superadditivity due to our nonorthogonal alphabet ; this is the simple difference between the two - shot rate and the one - shot rate .
we plot @xmath74 in figure 4 .
it has been suggested in ref .
@xcite that the differences @xmath75 can help define various notions of when two quantum states are most `` quantum '' with respect to each other ( and hence least `` classical '' ) .
when one goes to the limit @xmath76 one finds a well - behaved notion : two states are most quantum with respect to each other when they are @xmath77 apart .
figure 4 seems to indicate that @xmath74 plays no such simple role : at the very least , it means that this difference does a poor job of ferreting out the quantumness of two states in the geometrical sense already supplied by hilbert space .
let us now focus on the case we are most interested in for our experimental proposal : two very low photon - number coherent states @xmath19 and @xmath20 of a particular field mode .
we choose @xmath22 real so that the mean photon number in that mode is @xmath78 .
for the angles for which we demonstrated superadditivity , i.e. , @xmath79 , this translates to a mean photon number less than 0.03 in each transmission . in this case , we are well warranted in making the approximation @xmath80 where @xmath81 and @xmath82 denote the zero- and single - photon states of the mode , respectively .
moreover , we have @xmath83 in order to keep track of the separate transmissions , we encode each transmission in a different mode . for our purposes
it is convenient to choose two orthogonal circular polarizations .
but of different polarization modes .
such an alphabet takes the form @xmath84 and @xmath85 , where the tensor product structure now reflects the fact that one is using two field modes for each single transmission .
however , this alphabet of states is even less orthogonal than the one defined in eq .
( [ sukkel ] ) . ] expanding the measurement basis in terms of the photon number states , we thus have @xmath86 subscripts in these equations refer to righthand and lefthand circularly polarized light , respectively .
the measurement basis above is , of course , orthonormal .
however , after optimizing over @xmath64 as in the previous section , one finds that the coefficient of each @xmath87 component turns out to be of order @xmath22 while the other terms are of order one . because one is free to choose any measurement basis ,
we choose to ignore the small @xmath87 term for each @xmath88 .
this new basis @xmath89 is close to the optimal basis @xmath88 but allows the great simplification of not having to worry about how to distinguish @xmath87 from @xmath90 and @xmath91 .
we may then focus on experiments based on the absorption of at most a single photon .
the final step for defining our measurement scheme is to re - orthogonalize the vectors @xmath89 .
a simple convenient technique for this is the one introduced in ref .
let @xmath92 then clearly the vectors @xmath93 form an orthonormal set .
it is this basis that we will use in the experimental proposal , the main point of interest about it being that it contains no two - photon contributions .
of course , the new basis can not be optimal for achieving the rate @xmath61 already calculated , but for small @xmath22 it becomes arbitrarily good .
in fact , it is already sufficient for demonstrating superadditivity for @xmath94 ( see fig .
we now turn to the task of realizing the measurement explored in the last section . to carry this out
, we need the ability to perform an entangled measurement on two wave packets at a time .
we can achieve this collective decoding by mapping the orthonormal measurement basis in eq .
( [ mambo ] ) onto a set of orthonormal superpositions of three sublevels of a single atom ( see figure 5 ) .
the basic idea is to first transfer the information from the propagating light fields to photons inside an optical cavity and subsequently map the information from the cavity field to a single atom inside that cavity .
this can be accomplished as follows : first , the atom is prepared in a ground state with @xmath95 by optical pumping .
the presence of a single @xmath96 polarized cavity photon is then more than sufficient to induce a raman transition to the @xmath97 state with the help of a @xmath98-polarized laser field ( in fact , the advances in cavity qed have increased the [ four ] atom - cavity coupling to such a large degree that the saturation photon number for optical transitions is very small@xcite ; in particular , for the @xmath99 transition in cesium it is only @xmath100@xcite ) .
similarly , the presence of a single @xmath101 photon will induce the transition to @xmath102 , while if no cavity photon is present , the atom will stay in @xmath95 .
thus , the measurement scheme is based on the mapping @xmath103s @xcite ) .
once this mapping has been performed , we no longer rely on cavity fields . in order to avoid any disturbing effects from the laser field on level @xmath95 in the absence of a cavity photon
, we require the transition @xmath104 to be forbidden , which is easily accomplished by choosing @xmath105 transitions .
for example , one might consider the following transition between hyperfine multiplets in cesium @xmath106 moreover , the frequency @xmath107 of the laser field is chosen such that we are on two - photon resonance with the @xmath108 raman transitions , but far off resonance with respect to the excited states .
therefore , the latter will not be populated and no further transitions from @xmath97 or @xmath102 will occur .
once the information has thus been transferred from the polarizations to the atom in the cavity , the measurement basis is an orthonormal _ superposition _ of the three relevant atomic ground states @xmath102 , @xmath95 , and @xmath97 .
making a measurement of a superposition of these states is far more difficult than measuring the states themselves .
therefore , we first apply a unitary operation that transforms the basis of eq .
( [ mambo ] ) into the physical measurement basis .
this operation can be performed by a series of at most 16 appropriately timed raman pulses @xcite . in general , for an @xmath109 level system , with @xmath109 even , the unitary evolution can be controlled with a sequence of @xmath110 pulses consisting of two distinct perturbations in an alternating sequence@xcite .
while this scheme is not optimal for @xmath111 , it does give an upper bound for the required number of pulses .
once this transformation of basis has been performed , the only remaining task is to measure the projection onto each of the three possible hyperfine levels of our physical measurement basis . to perform this measurement , a magnetic field
is turned on adiabatically , causing a splitting of the energy of these otherwise degenerate hyperfine levels .
next , we use the technique of optical shelving to make a measurement of the levels @xcite . with this technique ,
a raman pulse is applied to cause a transition from the @xmath97 state into a secondary state that can then be driven on resonance to yield a large number of photons .
if the atom fluoresces at the driven frequency , the measurement outcome is @xmath112 , and the measurement is finished .
otherwise , if no fluorescence is detected , the atom will not be affected by this driving laser and the process is then repeated for the @xmath95 and @xmath102 states .
finally , let us note that atoms can already be held in a cavity for times exceeding @xmath113s @xcite , which is nearly sufficient for the measurements and laser manipulations discussed to be performed .
in addition , further improvements on trapping atoms in cavities will relax the conditions on timing .
the ability to hold single atoms in a cavity for a sufficient period of time will open up a world of possibilities for the field of communication@xcite .
we thank h. j. kimble , r. legere , h. mabuchi , m. sasaki , p. w. shor , and j. a. smolin for helpful discussions .
we thank a. landahl for a careful reading of the manuscript .
this work was supported by the quic institute funded by darpa via the aro , by the onr , and by the nsf .
caf also acknowledges the support of a lee a. dubridge fellowship . c. m. caves and p. d. drummond , rev .
phys . * 66 * , 481 ( 1994 ) .
a. s. kholevo , prob .
* 9 * , 110 ( 1973 ) .
t. m. cover and j. a. thomas , _ elements of information theory _ , ( john wiley & sons , new york , 1991 ) .
a. s. kholevo , rep .
phys . , * 12 * , 273 ( 1977 ) . a. s. kholevo , prob
. inf .
* 15 * , 247 ( 1979 ) .
a. peres and w. k. wootters , phys .
lett . * 66 * , 1119 ( 1991 ) . c. h. bennett , d. p. divincenzo ,
c. a. fuchs , t. mor , e. rains , p. w. shor , j. a. smolin , and w. k. wootters , phys .
a * 59 * , 1070 ( 1999 ) . c. a. fuchs , to appear in _ quantum communication , computing , and measurement 2 _ , edited by p. kumar , g. m. dariano , and o. hirota ( plenum press , ny , 1998 ) ; see also quant - ph/9810032 .
p. hausladen , r. jozsa , b. schumacher , m. westmoreland , and w. k. wootters , phys .
a * 54 * , 1869 ( 1996 ) .
a. s. holevo , ieee trans .
* 44 * , 269 ( 1998 ) .
b. schumacher and m. d. westmoreland , phys .
a * 56 * , 131 ( 1997 ) .
m. sasaki , k. kato , m. izutsu , and o. hirota , phys .
a * 236 * , 1 ( 1997 ) .
m. sasaki , k. kato , m. izutsu , and o. hirota , phys .
a * 58 * , 146 ( 1998 ) .
m. sasaki , t. sasaki - usuda , m. izutsu , and o. hirota , phys .
rev.a * 58 * , 159 ( 1998 ) .
w. nagourney , j. sandberg , and h. g. dehmelt , phys .
, 2797 ( 1986 ) .
t. sauter , w. neuhauser , r. blatt , p. e. toschek , phys .
rev.lett . * 57 * , 1696 ( 1986 ) .
j. c. bergquist , r. g. hulet , w. m. itano , and d. j. wineland , phys .
lett . * 57 * , 1699 ( 1986 ) .
a. peres , _ quantum theory : concepts and methods _ , ( kluwer academic , dordrecht , 1993 ) .
l. b. levitin , in _ quantum communications and measurement _ , edited by v. p. belavkin , o. hirota , and r. l. hudson , ( plenum press , ny , 1995 ) . c. a. fuchs and c. m. caves , phys .
lett . * 73 * , 3047 ( 1994 ) .
j. a. smolin , private communication .
l. p. hughston , r. jozsa , and w. k. wootters , phys .
a * 183 * , 14 ( 1993 ) .
j. kimble , in _ cavity quantum electrodynamics _ , edited by p. berman ( academic press , san diego , 1994 ) .
c. j. hood , m. s. chapman , t. w. lynn , and h. j. kimble , phys.rev .
lett . * 80 * , 4157 ( 1998 ) .
h. mabuchi , j. ye , h. j. kimble , appl .
b , in press ( 1999 ) ; see also quant - ph/9805076 .
c. k. law and j. h. eberly , opt .
* 2 * , 368 ( 1998 ) .
g. harel , v. m. akulin , phys .
lett . * 82 * , 1 ( 1999 ) .
j. i. cirac , p. zoller , h. j. kimble , and h. mabuchi , phys .
rev.lett . * 78 * , 3221 ( 1997 ) .
s. j. van enk , j. i. cirac , and p. zoller , science * 279 * , 205 ( 1998 ) . s. j. van enk , j. i. cirac , and p. zoller , phys .
lett . , 4293 ( 1997 ) . | we demonstrate superadditivity in the communication capacity of a binary alphabet consisting of two nonorthogonal quantum states . for this scheme ,
collective decoding is performed two transmissions at a time .
this improves upon the previous schemes of sasaki _
et al_. [ phys .
rev .
a * 58 * , 146 ( 1998 ) ] where superadditivity was not achieved until a decoding of three or more transmissions at a time .
this places superadditivity within the regime of a near - term laboratory demonstration .
we propose an experimental test based upon an alphabet of low photon - number coherent states where the signal decoding is done with atomic state measurements on a single atom in a high - finesse optical cavity .
pacs number(s ) : 03.67.-a , 03.67.dd , 03.67.lx , 03.67.hk , 03.65.bz , 32.80.-t , 33.55.ad , 42.65.pc |
there is a long - standing interest in how the magnus force , acting on vortices in superfluids and superconductors , changes as the system moves away from the galilean - invariant ( gi ) limit .
indeed , experiments indicate that the effective magnus force is very small both in conventional bulk superconductors except for very clean ones and in `` discrete '' superconductors , formed by josephson - junction arrays ( jja).@xcite in the first instance , there is a convincing explanation for this smallness , based on the spectral flow of fermions at the vortex core.@xcite the spectral flow creates an additional force on the vortex that reduces the total , effective magnus force nearly to zero .
however , the second case has remained something of a mystery .
various explanations of the smallness of the magnus force in jja have been reviewed in ref . .
one proposal is that in this case the magnus force is proportional not to the total density of electrons , but only to the `` offset charges '' , given by the deviation of the system from electrical neutrality.@xcite another proposal is that the effective magnus force vanishes exactly as a consequence of the particle - hole symmetry.@xcite however , volovik @xcite has argued that the particle - hole symmetry in these systems is not exact , and as a result the effective magnus force is nonzero , although small .
finally , we mention that when the josephson barrier is metallic , cancellation of the magnus force can be explained by a spectral flow mechanism similar to that in bulk superconductors .
this , however , does not work when the barrier is insulating .
there is a common theme to the above proposals : they all make use of specific properties of the electronic spectra or , alternatively , of the particle - hole symmetry already at the level of an effective description in terms of phases and charges of superconducting islands.@xcite as we will see below , that symmetry , present in the simplest model of jja , results in fact from neglecting the coupling between phase gradients and density fluctuations .
the question then is whether this assumption indeed applies in the discrete limit , or a nonzero magnus force persists no matter how discrete the system becomes . in this paper , we report results of a numerical study of the magnus force .
these results have been obtained by numerical solution of the classical gross - pitaevskii ( gp ) equation in two dimensions ( 2d ) .
since the classical approximation neglects the commutator of the bose fields @xmath0 , @xmath1 in comparison with the average density @xmath2 , it requires that the number of particles per site be large enough .
more precisely , the classical limit in jja is reached when the josephson energy is much larger than the charging energy.@xcite an equivalent condition is a^2 , [ limit ] where @xmath3 is the lattice spacing ( we assume a square or a nearly square lattice ) , and @xmath4 is the `` healing '' length , defined below .
the right - hand side of ( [ limit ] ) is a dimensionless measure of the interaction strength . to explore the role of discreteness ,
we solve the time - dependent gp equation on uniform spatial lattices with different values of the lattice spacing @xmath3 .
the relevant length scale to which @xmath3 can be compared is the `` healing '' length @xmath4 .
when @xmath5 , we reach the nearly gi limit , in which the gp equation describes a quasi - continuous neutral superfluid .
vortices have a core of size @xmath4 , which is resolved by the lattice . in the opposite limit , @xmath6
, vortices have no core , in the sense that there is no significant depletion of density anywhere .
the lattice sites can then be thought of as corresponding to individual islands , each of which is characterized by a value of the phase variable a model of a jja . more precisely , the lagrangian of our model in rescaled variables , for the case of a square lattice , is l = _ j ^_j ( i _ t - |_j|^2 + const ) _ j - 1a^2 _ ( ij ) |_i - _ j|^2 , [ l ] where the first sum is over all lattice sites , and the second over all nearest neighbors . note that if we write @xmath7 and neglect fluctuation of density @xmath8 in the second ( gradient ) sum in eq .
( [ l ] ) , the classical equations of motion become invariant under the transformation @xmath9 , @xmath10 . this is the particle - hole symmetry that was used in ref .
to argue the absence of magnus force in jja . here , we study the complete lagrangian ( [ l ] ) , including the coupling between @xmath11 and the phase gradients . on a coarse lattice , and in the presence of superflow
, the rotational invariance is broken so strongly that definition of the magnus force becomes a non - trivial matter .
indeed , we have found that in general vortices and anti - vortices do not even move symmetrically with respect to the superflow . however , when the superflow is along one of the lattice s main periods , they do , and we concentrate on this case in what follows . we have found that while on fine lattices vortices move with the superflow , as expected in the gi limit , on coarse lattices they move perpendicular to it .
we interpret this as vanishing of the effective magnus force in the discrete limit , in agreement with the experiments on jja.@xcite the paper is organized as follows .
we describe details of the numerical procedure in sect .
[ sect : method ] . in sect .
[ sect : pheno ] , we describe a phenomenological model that we use to interpret our numerical results . this model allows us to convert measurements of the longitudinal and transverse velocities of a vortex into values of the effective magnus and drag force coefficients .
numerical results are presented in sect .
[ sect : results ] .
[ sect : concl ] is a conclusion .
the continuum gp equation has the form i = - ^2 + g ||^2 , [ gp ] where @xmath0 is a complex scalar , the order parameter of the superfluid , @xmath12 is the mass of a fluid particle ( in units where @xmath13 ) , and @xmath14 is the interaction constant . for computational purposes , it is convenient to scale out the parameters by expressing length , time , and the order parameter in their `` natural '' units .
a natural unit of length is the `` healing '' length = ( 2 mg)^-1/2 , [ xi ] where @xmath15 is the average density of the fluid .
then , new , tilded variables are defined by the following relations : x & = & , [ tx ] + t & = & , [ tt ] + & = & , [ tpsi ] . in the new variables , the gp equation simplifies to i = - ^2 + ||^2 .
[ tgp ] note that by virtue of ( [ tpsi ] ) the rescaled average density is always equal to one : _
ave = = 1 .
[ tn ] to study the motion of vortices , we discretize eq .
( [ tgp ] ) on a uniform spatial lattice and solve it as an initial value problem , i.e. , knowing the state of the system at some initial moment , we calculate the state at later times .
this requires imposing suitable initial and boundary conditions ( see below ) .
the lattice in general has different lattice spacings in the @xmath16 and @xmath17 directions .
we used an operator - splitting algorithm with separate updates for the laplacian and potential terms in ( [ tgp ] ) .
updates corresponding to the laplacian were done using the crank - nicholson scheme , which is unconditionally stable .
the complete algorithm is unitary and second - order accurate in space and time . to avoid effects of the boundaries on the motion of the vortices , we use periodic boundary conditions in both directions .
the initial states for the runs are created in the following way .
we begin with the following field , containing a vortex and an anti - vortex ( the presence of an anti - vortex is necessary to satisfy the boundary conditions ) : ( z ) = _ z_+ _ z_- , where @xmath18 and @xmath19 are the desired ( complex ) positions of the vortex and anti - vortex , and @xmath20 $ ] denotes the set of positions including @xmath21 and a few mirror images with respect to the boundary .
then , evolving the system in the imaginary time , we cool the system down .
positions of the vortices during the cooling do not change , so we can place the vortices in convenient locations . to minimize effects of the vortex - anti - vortex interaction , we place them half of the total lattice length apart .
after that , we turn on a superflow and begin evolution in real time . a representative initial state , before and after the superflow was turned on , can be seen in figs .
[ woflow ] , [ wflow ] .
the aim of the simulations is to observe the motion of vortices in the presence of superflow .
the order parameter of the superfluid , being a complex scalar , can be written as = e^i , where @xmath22 is the local density of the superfluid .
then , @xmath23 is the local superfluid velocity .
the average velocity of the superflow @xmath24 is calculated as an average over the entire lattice .
the velocity of a vortex @xmath25 , on the other hand , is obtained from direct tracking of the vortex position during the simulation . in a gi system , we expect vortices to move with the flow : @xmath26
. if the invariance is broken , they may behave differently . to see the actual behavior , we break gi by solving the problem on increasingly coarser lattices .
while the vortex velocity can be measured directly in our simulations , converting these measurements into a value of the magnus force requires a model of forces acting on the vortex .
a fairly conventional model is available for an isotropic fluid ( which we expect to apply also in the gi limit on a lattice ) , but on coarse lattices modifications are needed . in this section , we review the conventional model , and then describe new effects introduced by the lattice .
the conventional ( phenomenological ) model includes three forces acting on a vortex , see fig .
[ forcediagram ] .
first , there is a drag force @xmath27 , directed against the vortex velocity .
it accounts for longitudinal momentum transfer from the vortex to the lattice and to excitations ( quasiparticles ) .
the latter channel is dissipative , i.e. , the work done by the drag force goes into excitation of the quasiparticle subsystem .
close to the gi limit , we expect the classical momentum transfer to quasiparticles to be ineffective , and hence the drag to be small .
second , there is an effective magnus force @xmath28 , perpendicular to the vortex velocity .
finally , there is a force , perpendicular to the superflow velocity and accounting for the work done by vortices as they unwind the superflow .
we refer to it as the lorentz force , @xmath29 .
perpendicular to the superflow velocity @xmath24 , @xmath30 perpendicular to the vortex velocity @xmath25 , and @xmath31 opposite to @xmath25 .
in dynamic equilibrium the sum of all forces vanishes , and the vortex moves at an angle @xmath32 with respect to the superflow . ] the coefficients @xmath33 , @xmath34 , and @xmath34 refer to unit inertial mass .
we will never need to discuss the actual value of the vortex inertial mass in this paper .
so , the equation of motion for the vortex is & = & _ drag + _ magnus + _ lorentz + & = & - - _ v + _ u .
[ eqm ] in complex notation , where we identify the @xmath16-direction with the real axis and the @xmath17-direction with the imaginary axis , the equation becomes = -v + u , where @xmath35 and @xmath36 .
the solution is easily found to be v = u + c ( -t ) , [ sol ] where @xmath37 is an integration constant . the exponential term is a transient that rapidly decays and turns out to be too small to be observed even at small times . at large times
, it drops out altogether .
then , the solution becomes a motion with a constant velocity at an angle @xmath38 , given by _
v = = /_v , with respect to the supercurrent .
the angle @xmath39 is related to the hall angle @xmath40 frequently used in the literature by @xmath41 .
the conclusion that vortices move in straight lines is well born out numerically .
notice that the steady velocity @xmath42 depends only on the ratios @xmath43 and @xmath44 , characterizing the magnus and drag forces .
measuring two components of the steady velocity , we obtain two equations for these two ratios , which can be solved with the result = , [ sigma_v ] + = .
[ eta ] an immediate consequence of the above expressions is that changing the sign of vorticity , i.e. , the signs of the coefficients @xmath34 and @xmath45 , changes the motion of a vortex ( which now becomes an anti - vortex ) in such a way that the projection of the vortex velocity on the direction of @xmath24 remains the same , while the orthogonal projection changes sign . in other words , a vortex and an anti - vortex move symmetrically with respect to the superflow . in general , for coarse lattices and superflow that is not parallel to one of the main periods of the lattice , we have found that the motion does not have that property .
we interpret this as a result of anisotropy introduced by the lattice and by the superflow direction . to account for anisotropy , the net force in eq .
( [ eqm ] ) needs to be replaced by = -m - _ v + _ u , where @xmath46 is a symmetric matrix that can depend on the direction of @xmath24 .
such a matrix has three independent elements , which now replace the single drag coefficient of the isotropic model . on the other hand ,
if the superflow velocity is along one of the main periods of the lattice , the vortex and anti - vortex do move symmetrically with respect to it . in this case
, we can introduce the effective magnus force and drag coefficients that are _ defined _ by eqs .
( [ sigma_v ] ) and ( [ eta ] ) . in what follows
, the superflow is always oriented along the @xmath16 direction , @xmath47 , and we present two types of results : one type is the ratios @xmath48 and @xmath49 themselves , which are directly measurable quantities , and the other is the effective force coefficients computed from eqs .
( [ sigma_v ] ) and ( [ eta ] ) .
simulations with different total lattice lengths have been carried out , with similar outcomes .
the ratio of the vortex velocity to the superflow velocity for square lattices of the same length 600 , ( in the rescaled length units ) , constant superflow velocity @xmath50 , and different lattice spacings is shown in fig .
[ velocities600 ] . because the vortex has to overcome pinning in the lattice cells , it moves by detectable jumps .
the data points were obtained by averaging the measured vortex velocities over long time intervals that begin some time after the start of the simulation .
we see that when the lattice spacing is close to 1 , in our dimensionless units , we obtain results expected for the gi limit : @xmath51 and @xmath52 , i.e. , @xmath53 , which means that vortices go with the flow . for large spacings , i.e. , in the discrete limit , the behavior changes radically .
now , @xmath54 and @xmath55 , i.e. , vortices move perpendicular to the current . between the two limits
, there is an interesting regime when @xmath56 , corresponding to motion with a hall angle of @xmath57 .
let us see how these results are reflected in the parameters of our phenomenological model . as discussed in subsect .
[ subsect : lat ] , we determine the effective force coefficients using eqs .
( [ sigma_v ] ) , ( [ eta ] ) .
the result is shown in fig .
[ forces600 ] . in the quasi - continuous limit ,
we obtain @xmath58 and @xmath59 , meaning that the magnus force coefficient is nearly equal to the lorentz force coefficient , while the drag coefficient is very small in comparison .
in contrast , in the discrete limit , the magnus force is nearly absent .
this is our main result : we have observed a gradual vanishing of the magnus force as we go from the continuous limit to the discrete one .
we note that , while the magnus force changes gradually , the drag coefficient has a rather sharp peak at the beginning of the region where @xmath60 , i.e. , when the vortex begins to move at @xmath57 to the flow . the simplest intuitive picture that might account for the vanishing of the magnus force is as follows .
as the lattice spacing increases , the depletion of the density in the region of large phase gradients [ the @xmath61 coupling discussed after eq .
( [ l ] ) ] becomes smaller , so that the system approaches the particle - hole - symmetric limit , in which the magnus force is absent.@xcite a small remaining force , caused by deviations from this ideal limit , can be overcome by a force exerted by the lattice , resulting in zero net magnus force . to make sure that the variations of the observables with the lattice spacing are not due to variation in the population of short - wavelength modes
, we check the power spectra .
[ powerspec ] shows the power spectrum of the field @xmath62 in the middle of the simulation for three different values of the lattice spacing . from the plot
, we infer that there are no major differences in the power spectra .
it is interesting to further explore the surprisingly broad intermediate range of lattice spacing where , as seen in fig .
[ velocities600 ] , the hall angle is close to @xmath57 .
we have found that this preferred direction is related to the geometry of the lattice , i.e. , it remains along the diagonal of the unit cell even after we go from the square unit cell to a rectangular one ( see fig .
[ anisotrop ] ) .
the main result of the present paper is that in the classical limit ( [ limit ] ) , vortices in superfluids on coarse ( but uniform ) lattices , in the presence of a superflow parallel to one of the main periods , move perpendicular to the superflow .
we interpret this result as a transition from the full magnus force in the galilean - invariant limit to vanishing effective magnus force in the discrete limit , in agreement with the observed smallness of the magnus force in jja.@xcite our results are based on direct numerical simulations of the discrete superfluid ( [ l ] ) and do not assume a priori any symmetry that might prohibit the magnus force in the discrete limit .
another potential application of our results is vortex motion in cold atomic gases confined in optical lattices .
if a sufficient degree of experimental control over parameters in either of these systems can be reached , intermediate points in the transition from the galilean - invariant limit to the discrete limit may become observable .
notably , in our simulations these intermediate points include a somewhat counterintuitive regime where , for a broad range of lattice spacings , the average vortex velocity is along the diagonal of the unit cell .
99 j. m. harris , y. f. yan , o. k. c. tsui , y. matsuda , and n. p. ong , phys .
* 73 * , 1711 ( 1994 ) . for a review , see r. fazio and h. van der zant , phys .
rep . * 355 * , 235 ( 2001 ) .
n. b. kopnin and v. e. kravtsov , zh .
1644 ( 1976 ) [ jetp * 44 * , 861 ( 1976 ) ] . g. e. volovik , zh .
* 104 * , 3070 ( 1993 ) [ jetp * 77 * , 435 ( 1993 ) ] .
a. van otterlo , m. feigelman , v. geshkenbein , and g. blatter , phys .
lett . * 75 * , 3736 ( 1995 ) .
m. v. feigelman , v. b. geshkenbein , a. i. larkin , and v. m. vinokur , pisma zhetf * 62 * , 811 ( 1995 ) [ jetp lett . * 62 * , 834 ( 1995 ) ] .
m. stone , phys .
b * 54 * , 13222 ( 1996 ) ; g.e .
volovik , cond - mat/9707136 .
r. fazio , a. van otterlo , g. schn , h. s. j. van der zant , and j. e. mooij , helvetica physica acta * 65 * , 228 ( 1992 ) .
sonin , phys .
b * 55 * , 485 ( 1997 ) .
yu.g . makhlin and g.e .
volovik , pisma zh .
. fiz . * 62 * , 923 ( 1995 ) [ jetp lett . * 62 * , 941 ( 1995 ) ] . | motion of vortices in two - dimensional superfluids in the classical limit is studied by solving the gross - pitaevskii equation numerically on a uniform lattice .
we find that , in the presence of a superflow directed along one of the main lattice periods , vortices move with the superflow on fine lattices but perpendicular to it on coarse ones .
we interpret this result as a transition from the full magnus force in the galilean - invariant limit to vanishing effective magnus force in a discrete system , in agreement with the existing experiments on vortex motion in josephson junction arrays . |
materials which exhibit a metal - insulator transition concomitant with a structural phase transition naturally raise questions about the driving mechanism of the phase transition , i.e. if it is possible to single out what is the role played by the electronic or lattice degrees of freedom and how they affect the electronic structure of such materials . in this context
, studies of the non - equilibrium dynamics have opened a new dimension in the investigation of correlated materials in recent years .
time- and angle - resolved photoemission spectroscopy ( trarpes ) has proved to be a major technique , giving access to the momentum - resolved electronic band structure of solids out of equilibrium @xcite . in the case of charge density wave ( cdw ) materials
, the most direct approach follows the photoinduced changes of the cdw spectral features , namely the closing of the cdw gap or the disappearance of the cdw backfolded bands directly in the time - domain .
many studies have focussed on the first hundred femtoseconds following the photoexcitation , in order to figure out whether these changes occur faster or not than the lattice motion related to the structural distortion , the timing of which is given by the period of the relevant phonon mode .
the recovery dynamics of cdw order have received less attention , however we will show below that they also provide fundamental information on the role of electronic and phonon degrees of freedom .
cdw materials undergo a low temperature phase transition resulting in a new translational periodicity in the electronic density ( the cdw ) , as well as a periodic lattice distortion ( pld ) .
one of these materials , 1_t_-tise@xmath0 , undergoes a peculiar second - order cdw transition@xcite below @xmath1 k. the mechanism of its cdw transition has been debated intensively over the last decades , in terms of either an electronic @xcite or a structural origin @xcite .
time - resolved techniques seem to give a decisive answer by focussing on the ultrafast melting of the cdw and pld and disentangling the roles of electrons and lattice on the femtosecond timescale @xcite .
these studies are consistent with a purely electronic mechanism of the cdw transition in 1_t_-tise@xmath0 , which might be an instability towards an excitonic insulator phase . in such a scenario , electrons and holes from distinct low - energy bands ,
are bound together in excitons by the interband coulomb interaction and their macroscopic spontaneous condensation at low temperature leads to a new ground state @xcite .
a recent ultrafast terahertz spectroscopy study@xcite stressed the importance of the lattice degree of freedom , which is shown to stabilize the cdw phase together with the excitonic mechanism . in parallel , a few theoretical models@xcite have demonstrated that the cdw ( and its concomitant pld ) can be cooperatively stabilized by both the interband coulomb interaction and the electron - phonon coupling . in the light of these recent studies , what is then
the role of the excitonic mechanism in 1_t_-tise@xmath0 ? does is alone provoke the cdw instability or does it merely assist the lattice in driving the cdw transition ? here
, in order to address these questions , we investigate the femtosecond dynamics of both the recovery of the cdw phase in 1_t_-tise@xmath0 at 30 k , as well as of high - temperature cdw fluctuations at 240 k , which precedes the cdw instability .
we exploit the full capabilities of trarpes by looking at the energy- and momentum - resolved recovery dynamics of the cdw correlations . by doing
so we unveil the dynamics of both the electron and lattice degrees of freedom .
below @xmath2 , in the cdw phase , we observe the coherent oscillations of the cdw amplitude mode for the first time with trarpes .
above @xmath2 , we single out a fast component in the recovery of the cdw correlations with a relaxation time @xmath3 100 fs . we interpret it as the consequence of the ultrafast recovery from the screening of the coulomb interaction responsible for the electron - hole correlations , which precedes the cdw phase transition .
the observation of these two contributions - the coherent oscillations of the cdw amplitude mode at low temperature and the ultrafast electronic recovery at high temperature - supports the idea that both the interband coulomb interaction and the electron - phonon coupling cooperatively stabilize the cdw phase ( and the concomitant pld ) in 1_t_-tise@xmath0 .
the paper is divided into two main parts : in part [ sec_lt ] , we study the low temperature ( 30 k ) cdw phase using a mid - infrared excitation .
we show how the excitation generates a direct optical transition between the backfolded valence and conduction bands in section [ subsec_lt_res ] .
the resulting coherent oscillation of the amplitude mode is then analyzed and explained as the result of a bonding - antibonding optical transition in section [ subsec_lt_phonon ] .
part [ sec_rt ] shows the results obtained for the cdw fluctuation regime at 240 k using a near - infrared excitation .
we demonstrate that the excitation generates a direct transition at the center of the brillouin zone in section [ subsec_rt_res ] and subsequently induces a change in the spectral function of the conduction band ( section [ subsec_rt_sw ] ) .
the transient electronic temperature is extracted in section [ subsec_rt_te ] . in section [ sec_disc ] ,
we discuss all these results and their interpretation within the framework of the cdw phase and fluctations in 1_t_-tise@xmath0 .
we conclude our study in section [ sec_conclusion ] .
fig . [ fig_1 ]
( a ) presents the schematic electronic band structure of 1_t_-tise@xmath0 near the fermi level @xmath4 in the cdw phase .
it consists of a valence band ( vb ) at @xmath5 , the center of the brillouin zone , and a conduction band ( cb ) at @xmath6 , the border of the brillouin zone ( see fig . [
fig_1](e ) for the corresponding hexagonal brillouin zone ) .
these two bands are the main components of the low energy band structure at room temperature . below 200 k , in the cdw phase
, a replica of the valence band , the cdw band , appears at @xmath6 as a direct consequence of the new periodicity of the electronic density in the cdw phase , see fig .
[ fig_1 ] ( a ) ( dashed line ) .
in addition , replicas ( at symmetry - equivalent @xmath6 point ) of the cbs appear above @xmath4@xcite .
our pump - probe measurements were obtained with the artemis time - resolved arpes endstation at the central laser facility of the rutherford appleton laboratory @xcite .
1_t_-tise@xmath0 samples were cleaved in a base pressure of @xmath7 mbar and were excited by 100 fs @xmath8-polarized laser pulses both in the mid - infrared ( mir ) region , at 3100 nm ( 400 mev ) , and in the near - infrared ( nir ) region , at 1400 nm ( 890 mev ) .
the sample was probed by 30 fs @xmath9-polarized laser pulses with 21 ev photons .
the total energy resolution was about 300 mev . in fig .
[ fig_1 ] , we show trarpes data measured in the cdw phase at 30 k when exciting the sample with mir pump pulses at 3100 nm with a fluence of 0.4 mj/@xmath10 . at this wavelength , electrons are excited resonantly at @xmath6 between the backfolded valence band and the backfolded cb ( see the orange arrows on fig .
[ fig_1 ] ( a ) ) .
[ fig_1 ] ( b ) and ( c ) show trarpes intensity plots taken before the arrival of the pump pulse , @xmath11 fs ( i.e. at thermal equilibrium ) and after the arrival of the pump pulse , @xmath12 fs , at which time the pump - induced effects are the strongest . at 30 k
, 1_t_-tise@xmath0 is in the cdw phase , meaning that a well - developed cdw band is measured at @xmath6 below the cb .
however , due to the limited energy resolution of our experiment , these two bands can not be resolved fig .
[ fig_1 ] ( b ) and ( c ) and the cb can hardly be distinguished .
is very intense ( see supplementary materials@xcite ) and dominates over the cb , so that it dominates also the spectral weight distribution when convolved by the experimental energy resolution . ] nevertheless , comparing the spectra measured at @xmath12 fs with those at equilibrium , one distinguishes clear differences , which are more obvious in the corresponding difference image in red / blue false colour plot .
such a difference intensity plot is shown in fig .
[ fig_1 ] ( d ) , for @xmath12 fs , giving a clear view of the dynamics in our trarpes data .
it emphasizes the main dynamical changes , namely the disappearance of the cdw band ( in blue ) and the accumulation of excited electrons in the cb ( in red ) .
the resonant mir excitation ( 400 mev ) taking place at @xmath6 also transfers population from the top of the cdw band into the unoccupied states .
this can be seen directly by looking at the time - dependent photoemission intensity integrated in the different energy versus momentum regions depicted by boxes in fig .
[ fig_1 ] ( d ) .
the energy position of such boxes is also indicated in fig .
[ fig_1 ] ( f ) , where we plot energy distribution curves ( edcs ) taken before ( black ) and 50 fs after photoexcitation ( red ) . here
the photoinduced changes below the fermi level @xmath4 are mostly due to the change of the spectral function of the cdw band . however , part of it is also due to the resonant optical transition taking place at @xmath6 with the mir pump pulse , as schematically displayed in fig .
this can be observed in the transient photoemission intensity curves shown in fig .
[ fig_1 ] ( g ) , integrated from the corresponding boxes in graph ( d ) . from box 1 ( blue curve ) , taken near @xmath4 , an oscillation is observed and will be discussed in the next section . at time 0 , an additional sharp peak can be distinguished , which is coming from the depopulation of the cdw band due to the pump pulse .
there is correspondingly a peak at about 0.4 ev higher in energy , in the unoccupied states .
this can be seen in the time - resolved intensity of box 2 ( red curve ) where we observe a transient population of a ( backfolded ) conduction band at @xmath6 .
this indicates that at time 0 , electrons are transferred from occupied to unoccupied states at @xmath6 , i.e. that an optical transition takes place at @xmath6 .
finally , the green curve shows the intensity integrated in box 3 , which is the region where the reduction of the photoemission intensity is the strongest .
however , we avoid this region in our further analysis at low temperature , as the nature of the transient signal here is not clear . indeed , due to the significant energy resolution of our experiment ( about 300 mev ) , the signal coming from the cdw band can not be well separated from the one from the conduction band at @xmath6 exactly ( see also the supplementary materials for a high - energy resolution spectrum taken at @xmath6 exactly@xcite ) . ( a ) photoemisson intensity map in false color showing the photoemission intensity difference map taken at 30 k for a mir pump pulses ( same as in fig .
[ fig_1 ] ( d ) ) .
( b ) absolute value of the normalized photoemission intensity changes integrated in the black boxes of graph ( a ) , together with the fit of the coherent phonon oscillation ( total fit , green curve ) and the exponential part of the fit only ( dashed blue curve ) . ( c )
schematic electronic structure of 1_t_-tise@xmath0 in the cdw phase at @xmath6 .
( d ) schematic of the displacements involved by the @xmath13 coherent phonon excitation by the mir pump pulse for a tise@xmath14 octahedron .
, width=321 ] we now look in more detail at the recovery dynamics of the spectral features associated with the excited electrons and with the cdw . the ( normalized ) time - dependent photoemission intensity of these features is shown in fig .
[ fig_2 ] ( b ) .
the melting dynamics of the cdw band has been already extensively studied and shown to be faster in time for larger pump fluences , an effect assigned to the screening by the transient generation of free charge carriers and therefore consistent with a purely electronic mechanism .
@xcite . here , these transient intensities have been integrated from the regions shown by boxes in fig .
[ fig_2 ] ( a ) , corresponding to the excited electrons in the cb ( red curve with square markers ) and the bottom of the backfolded vb ( blue curve with discs ) .
they both show an ultrafast response which recovers mostly within 2 ps followed by a slower recovery which is not over at 9 ps .
most interesting are the dynamics within 2 ps .
we emphasize that for states at binding energies larger than 400 mev ( the pump photon energy ) there are no unoccupied final states available for direct photoexcitation .
the dynamics observed in the cdw band is therefore due to a photoinduced change of its spectral function .
the photoemission intensity integrated at the bottom of the backfolded vb displays a clear oscillation on top of an exponential decay , up to about 1 ps .
the oscillation can be well described by a cosine with a period of about 290 fs , which correspond to 3.4 thz , and a phase of about @xmath15 .
these numbers are obtained by fitting the transient photoemission intensity of the coherent phonon oscillation using the following function @xmath16 this represents a sum of two incoherent exponential components , caracterized by two amplitudes , @xmath17 and @xmath18 , with two relaxation times , @xmath19 and @xmath20 , together with an oscillatory component , characterized by an amplitude , @xmath21 , a frequency , @xmath22 , a phase , @xmath23 , and damped by an exponential of damping time @xmath24 .
the two incoherent exponential components are necessary to achieve the best fit .
this function is then convolved with a gaussian having the experimental time resolution as a full width at half maximum ( 100 fs ) .
the following values give the best fit .
for the relative amplitudes , @xmath25 , @xmath26 and @xmath27 .
for the relaxation times , @xmath28 fs , @xmath29 fs and @xmath30 fs . the phonon period is @xmath31 fs and the phase is 0 .
this mode has already been observed in time - resolved reflectivity data from 1_t_-tise@xmath0 @xcite and was attributed to the @xmath13 amplitude phonon mode of the cdw @xcite , corresponding to a breathing mode of a tise@xmath14 octahedron , as depicted in fig .
[ fig_2 ] ( d ) @xcite .
this oscillation is highly localized in momentum space and is absent from the dynamics of the excited electrons .
such a coherent phonon oscillation is caused by the excitation of coherent ion motions by the pump laser pulse .
the 0 phase of a cosine - like oscillation indicates that it is a displacive excitation , meaning that the potential energy surface of the corresponding ions , which is suddenly modified by the electronic excitation , triggers the coherent phonon oscillation . in fig .
[ fig_2 ] ( b ) , the dynamics of the excited electrons and that of the exponential decay of the cdw band are very similar ( compare the exponential decay fit of the cdw band intensity ( dashed blue curve ) with the one of the hot electrons in fig .
[ fig_2 ] ( b ) ) , supporting the idea that the transient change of the potential energy surface is modified by the excited electron distribution .
( a ) photoemisson intensity map in false color showing the vb along @xmath32 , near @xmath5 , measured at 30 k before time 0 with a 1300 nm pump photon wavelength .
( b ) corresponding false color plots showing the intensity difference induced by the pump pulse ( positive and negative intensity changes in red and in blue , respectively ) near @xmath5 . electrons and holes are excited resonantly at time 0 by the pump pulse , the energy of which is shown by the black arrow .
( c , d ) corresponding intensity maps along @xmath33 , near @xmath5 , where part of the conduction band and the cdw band can be seen on the right side of the graphs . ( e )
hexagonal brillouin zone of 1_t_-tise@xmath0 , with the two directions relevant to graphs ( a , b , c , d ) .
( f ) photoemission intensity traces as a function ot time delay integrated in the corresponding numbered black boxes on graphs ( b , d ) .
the curves are normalized to their maximum .
, width=321 ] in the cdw phase , it has been shown that the electronic structure of 1_t_-tise@xmath0 can be well described by a mean - field hamiltonian describing one hole - like band at @xmath5 coupled to three symmetry equivalent electron - like bands near @xmath6 ( see the corresponding schematical fermi surface in fig .
[ fig_1 ] ( e ) ) @xcite . at @xmath6 ,
the three conduction bands are degenerate ( without any coupling ) and the effective hamiltonian has the form @xmath34 where the order parameter @xmath35 describes the strength of the coupling between the single valence band @xmath36 and the three symmetry equivalent conduction bands @xmath37 ( @xmath38 ) , in the basis of the wave functions for the corresponding electronic states , namely @xmath39 .
once the coupling between the valence and conduction bands takes a finite value , the valence band gets backfolded to @xmath6 and the conduction bands get backfolded to @xmath5 and other @xmath6 points .
the new band dispersions along the high - symmetry direction @xmath33 are shown in fig .
[ fig_2](c ) ( for realistic dispersions fitted from the experiment@xcite ) . for the extrema of the valence and conduction bands
, we use the values @xmath40 mev , @xmath41 mev at @xmath5 and @xmath6 , respectively . for the order parameter , we use @xmath42 mev , which fits well with the experiment at low temperature . diagonalizing the hamiltonian @xmath43 leads to the following eigenvalues at @xmath6 @xmath44 the values @xmath45 and @xmath46 correspond to the backfolded valence and backfolded conduction bands ( at @xmath6 ) . the mid - infrared pump photon energy of 400 mev matches approximatively the resonance between these two bands , which can be defined as the cdw gap in 1_t_-tise@xmath0 .
the eigenvectors corresponding to these eigenvalues are @xmath47 as a consequence , @xmath45 corresponds to the antibonding state and @xmath46 to the bonding state . in 1_t_-tise@xmath0 , the vb at @xmath5 is mostly of se @xmath48 character and the cb at @xmath6 of ti @xmath49 character . as shown above , in the cdw phase
, these bands will effectively hybridize due to their coupling and have a mixed character .
we have just demonstrated that the states at the top of backfolded vb band have a bonding symmetry , while those at the bottom of the backfolded cb have an antibonding symmetry .
this suggests how the resonant mir excitation triggers a displacive @xmath13 cdw amplitude mode by exciting electrons from bonding to antibonding states in 1_t_-tise@xmath0 ( see fig .
[ fig_2](d ) ) .
summarizing , in the low temperature phase of 1_t_-tise@xmath0 , the resonant mir excitation triggers a coherent lattice response of the cdw amplitude mode , observable in the recovery of the backfolded valence band of the material .
this differs from the case of higher pump photon energies @xcite for which no coherent oscillations are observed .
we explain it here by the fact that this resonant optical transition occurs between bonding and antibonding states ( at @xmath6 ) of the tise@xmath14 octahedron acting efficiently on the ionic position relevant to the cdw distortion and stimulating this way a displacive excitation .
we now turn to the high temperature trarpes data .
the phase transition in 1_t_-tise@xmath0 is continuous ( second - order ) @xcite and photoemission data@xcite have shown that short - range cdw fluctuations persist well above @xmath1 k. such fluctuations are known to lead to spectral features anticipating those of the low temperature cdw phase , e.g. with a cdw pseudo - gap@xcite . in the case of 1_t_-tise@xmath0 , no cdw band is observed , but a broad cb at @xmath6 ( see supplementary materials for a high - energy resolution arpes spectrum@xcite ) , the width of which results from the short - range cdw fluctuations .
we have therefore also measured 1_t_-tise@xmath0 at 240 k , above @xmath2 , to investigate the dynamics of these spectral features . at 240 k , due to the disappearance of the sharp cdw bands at high binding energies ,
a resonant mir excitation can not be used ( accordingly , the cdw peak in the optical conductivity vanishes @xcite ) .
we therefore used a shorter wavelength , 1300 nm , in the nir regime .
the trarpes data of fig .
[ fig_3 ] measured near @xmath5 , along @xmath32 ( see graph ( a ) ) , demonstrate that the nir pulses excite electrons by an optical transition taking place at around @xmath5 , as illustrated in fig .
[ fig_4 ] ( a ) .
this direction in the brillouin zone is depicted in fig .
[ fig_3 ] ( e ) .
shortly after time 0 ( at @xmath50 fs ) , one sees a clear depletion of intensity at the top of the valence band ( vb ) at @xmath5 and accordingly strong intensity above the vb , in a conduction band ( cb ) , which is slightly less than 1 ev above ( see graph ( b ) ) the top of the vb . the black arrow has a length of 890 mev , demonstrating this electronic resonance with the pump pulse . at later times , the excited hot electrons are scattered away from @xmath5 ( see graphs ( c ) and ( d ) ) and accumulate into the cb at @xmath6 . the time delay difference between these different events is shown in fig .
[ fig_3 ] ( f ) .
this behavior of the momentum path of the hot electrons in the unoccupied states is similar to what has been already observed by rohde _
@xcite for a much larger pump photon energy .
photoemission intensity maps taken at 240 k with nir excitation are shown in fig .
[ fig_4 ] ( b ) at negative ( @xmath51 fs ) and positive ( at @xmath52 fs ) time delays . at 240 k ,
i.e. above the critical temperature of the cdw phase , there is not clear cdw band at high binding energies anymore .
our trarpes data have been acquired with a pump photon wavelength of 1300 nm resonant with an optical transition at @xmath5 ( see previous section ) .
the difference image in red / blue false colour plot , fig .
[ fig_4 ] ( c ) , shows the accumulation of excited electrons in the cb ( in red ) and a decrease of photoemission intensity at the bottom of the cb ( in blue ) the transient photoemission intensities of the excited electrons and of the bottom of the cb are displayed in fig .
[ fig_4 ] ( d ) .
the corresponding time - dependent photoemission intensities integrated in the corresponding boxes in this graph are shown in fig .
[ fig_4 ] ( e ) .
interestingly the decrease of intensity at the bottom of the cb at @xmath6 ( in blue ) displays a dynamics with 2 timescales .
its recovery dynamics are therefore fitted with a bi - exponential function ( convoluted with a gaussian for experimental time resolution ) , involving a fast and a slow component , with decay constants @xmath53 fs and @xmath54 fs , respectively .
we argue below that this decrease of intensity is due to a photoinduced change of the spectral function . in fig .
[ fig_6 ] , we demonstrate that there is no resonant excitation taking place at @xmath6 at 240 k with a pump photon wavelength of 1300 nm ( 0.95 ev ) , in the nir regime .
graph ( b ) in fig .
[ fig_6 ] displays edcs taken at @xmath6 , before time 0 ( black ) and at 130 fs ( red ) .
graph ( c ) shows the transient photoemission intensity integrated in the boxes 1 and 2 in graph ( a ) .
box 1 is situated in a region where the transient suppression of the photoemission signal is the largest and box 2 is situated 0.95 ev higher in energy , in the unoccupied states .
the red curve from box 2 shows clearly that there is hardly any transient signal in this region , proving that there is no resonant excitation taking place at @xmath6 at 240 k for 1300 nm pump photon wavelength ( at variance with the case of mir pump photons at 30 k , see fig .
[ fig_1 ] ) .
we can therefore attribute the photemission intensity changes observed below @xmath4 to a photoinduced change of the spectral function of the conduction band , which we will explain further below in our discussion of the results .
( a ) photoemisson intensity map in false color showing the photoemission intensity difference map taken at 240 k ( same as in fig .
[ fig_4](c ) ) .
the blue arrows indicate the size of the pump photon energy ( 0.95 ev , 1300 nm ) in comparison to the energy axis .
( b ) edcs taken at @xmath6 before time 0 ( black curve ) and at 130 fs ( red curve ) .
( c ) absolute value of the normalized photoemission intensity changes integrated in the black boxes of graph ( a ) .
, width=321 ] we now focus on the dynamics measured in the cb at @xmath6 near the fermi edge .
we exploit here the fact that at this temperature , the cb in our sample is partially filled @xcite and analyse the slope of the corresponding fermi - dirac cut - off to monitor the transient electronic temperature , in a similar way to perfetti _ et al .
indeed , at @xmath6 , at 240 k , there is a flat blue - red region in the difference plot of fig .
[ fig_5 ] ( c ) ( see dashed box ) , typical from a heated metallic fermi - dirac edge .
the ( normalized ) edcs in this box are averaged over momentum and plotted in fig .
[ fig_5 ] ( a ) at different time delays , showing a clear increase of temperature .
they are fitted by an exponential function@xcite above @xmath4 , in order to avoid the changes of the spectral lineshape at lower binding energies ( coming from the photoinduced changes in the spectral function ) .
this is shown in fig .
[ fig_5 ] ( b ) . the extracted electronic temperature ( black curve with discs )
is displayed in fig .
[ fig_5 ] ( d ) ( after removing the contribution of the energy resolution@xcite about 300 mev ) . in the spirit of a two - temperature model , the electronic temperature decreases here in about 1 ps as the result of coupling to the lattice , the temperature of which should increase by about 40 k ( if heat diffusion is neglected ) for the pump fluence used here and the specific heat @xcite of 1_t_-tise@xmath0 .
( a ) edcs taken at @xmath6 at 240 k ( integrated inside the dashed box of graph ( c ) ) for different time delays .
the dashed grey line is the reference edc at -600 fs .
( b ) the same edcs fitted by an exponential .
( c ) false colour plot of the transient photoemission intensity at @xmath6 ( same as in fig .
[ fig_4](c ) ) .
( d ) comparison of the electronic temperature with the fitted transient photoemisson intensity at the bottom of the cb at @xmath6 ( black box in the blue region of fig .
[ fig_4 ] ( c ) ) and the corresponding slow component fitted from the spectral feature ( see main text ) .
, width=321 ]
the reduction of photoemission intensity observed at the bottom of the cb in fig . [ fig_4 ] ( c ) comes from a photoinduced change of its spectral function .
we interpret this effect as the suppression of the high binding energy tail of the cb , which is the spectral signature of the cdw fluctuations@xcite .
this means that the intensity of the cdw fluctuations are reduced by the pump pulse . in fig .
[ fig_4 ] ( e ) , we focus on the dynamics of these cdw fluctuations . the recovery dynamics are fitted with a bi - exponential function ( convoluted with a gaussian for experimental time resolution ) , involving a fast ( @xmath53 fs ) and a slow ( @xmath54 fs ) component . to understand these two timescales , we plot in fig .
[ fig_5 ] ( d ) the fitted time - dependent photoemission intensity at the bottom of the cb ( grey line ) together with the fitted slow component ( green ) and the extracted electronic temperature ( black ) .
this comparison suggests that the two timescales of the dynamics of the cdw fluctuations do not match the decay behavior of the electronic temperature ( especially the sharp peak near time 0 ) .
however , the slow component ( green ) timescale matches rather well the long timescale of the electronic temperature , indicating that part of the quenching of the cdw fluctuations is due to the transiently increased electronic temperature .
we interpret the fast component of the suppression of the cdw fluctuations as a consequence of the ultrafast screening of the coulomb interaction , which is enhanced by the new charge carriers created by the optical excitation@xcite .
the intensity of these fluctuations depends strongly on the strength of the interband coulomb interaction ( see supplementary materials for an analytical formula of the self - energy of the cb@xcite ) . in the framework of the excitonic insulator phase , this is the attractive interaction binding an electron and a hole in an exciton . in a previous study@xcite ,
the resulting self - energy leading to the broadened cb was derived and shown to have a particularly high value at the bottom of the cb .
this reveals then that electrons at the bottom of the cb experience a high scattering rate@xcite , naturally explaining the ultrafast recovery of the cdw fluctuations . in fig .
[ fig_4 ] ( f ) , we depict how this electron - hole interaction between the valence and conduction bands occurs . here
an electron jumps from the cb to the vb , scattering another electron from the vb to the cb .
this can effectively been seen as the scattering of an electron in the cb with a hole in the vb .
the strength of the cdw fluctuations depends strongly on the coulomb interaction and also on the electronic and lattice temperature@xcite .
therefore they tend to be suppressed not only by enhanced screening , but also by the increased transient electronic temperature .
the combination of these two observations explains the bi - exponential decay of the photoemission intensity suppression at the bottom of the cb ( see fig .
[ fig_4 ] ( e ) ) , namely the recovery of the cdw fluctuations .
we ascribe the fast component ( @xmath53 fs ) to the relaxation of the enhanced screening by the efficient electron - hole scattering and the slow component to a quasi - thermal recovery following the relaxation of the electronic temperature .
our interpretation that the fast component is due to the electron - hole correlations as a strong scattering channel supports their occurence at temperature above but near @xmath2 .
these correlations give rise to the cdw fluctuations in this temperature range@xcite and play an important role in stabilizing the low - temperature cdw phase . at 30 k ,
the situation is different as a well - formed pld is present .
the longer timescale dynamics of the cdw band display a clear response of the lattice degrees of freedom .
the @xmath13 cdw amplitude mode is observed as a coherent phonon oscillation , provided that the cdw band is selectively excited by a resonant mir photon pulse generating excited electrons with a minimal excess kinetic energy with respect to the cdw gap .
the fact that the intensity of the cdw band oscillates with the frequency of the @xmath13 cdw - related phonon mode implies that the order parameter of the cdw phase is strongly coupled to the corresponding atomic motion .
furthermore , this coupling is very localized in momentum space , namely only for states in the cdw band .
this means that both the excitonic interband coulomb interaction and the electron - phonon coupling contribute to the order parameter of the cdw phase @xcite .
in the first 100 fs , the melting of the cdw order is dominated by the suppression of the excitonic part , as observed by rohwer _
@xcite with rise times as fast as 20 fs for large fluences ( up to 5 mj/@xmath10 ) . in that sense ,
our experimental results support a model according to which strong electron - hole correlations , giving rise to short - range cdw fluctuations above @xmath2 , contribute to the stabilization of the cdw phase together with the electron - phonon coupling below @xmath2@xcite .
in conclusion , we have probed both the femtosecond dynamics of the cdw phase in 1_t_-tise@xmath0 at 30 k , as well as the dynamics of the high - temperature fluctuation regime at 240 k using trarpes . below @xmath2 , in the cdw phase , we observe the coherent oscillation of the cdw amplitude mode for the first time with trarpes using a mid - infrared excitation resonant with the cdw gap . above @xmath2 , we single out a fast component with a relaxation time @xmath3 100 fs . we interpret it as the consequence of the ultrafast screening of the interband coulomb interaction responsible for the electron - hole correlations .
we propose that these electron - hole correlations act as an efficient scattering channel in the relaxation of excited electrons . in this way
, we have unveiled the dynamics of both the electron and lattice degrees of freedom in 1_t_-tise@xmath0 .
the authors thank h. beck , p. aebi , k. rossnagel for useful discussions . c.m .
acknowledges support by the swiss national science foundation under grant number @xmath55 , as well as by the alexander von humboldt foundation .
access to artemis was provided by stfc . | using time- and angle - resolved photoemission spectroscopy with selective near- and mid - infrared photon excitations , we investigate the femtosecond dynamics of the charge density wave ( cdw ) phase in 1_t_-tise@xmath0 , as well as the dynamics of cdw fluctuations at 240 k. in the cdw phase , we observe the coherent oscillation of the cdw amplitude mode . at 240 k
, we single out an ultrafast component in the recovery of the cdw correlations , which we explain as the manifestation of electron - hole correlations .
our momentum - resolved study of femtosecond electron dynamics supports a mechanism for the cdw phase resulting from the cooperation between the interband coulomb interaction , the mechanism of excitonic insulator phase formation , and electron - phonon coupling . |
icosahedrally symmetric particles with local symmetries play a crucial role in virology , because viruses encapsulate their genome in protein containers , called viral capsids , that are organised in protein clusters centred on the symmetry axes of the capsids .
these axes of symmetry may be global , i.e. corresponding to rotations that leave the whole structure invariant and are described by the overall group structure , or local , in which case only the local environments around the symmetry axes .
an example of a local symmetry is the discrete rotation about the centre of the capsomere , that leaves only the capsomere itself , but not the whole capsid , invariant .
the classification of all possible local symmetries of icosahedral surface structures hence determines the spectrum of possible viral capsids with icosahedral symmetry .
viruses with icosahedrally symmetric capsids occur predominantly in nature @xcite . by making use of symmetry
, a virus can efficiently encode the structure of its capsid : it only needs to encode the locations of the protein subunits modulo symmetry , that is only the locations of a subset of the proteins , and the locations of all other proteins follow by symmetry . since the icosahedral group is the finite symmetry group of largest order in three dimensions ( order 60 , as opposed to order 48 for the octahedral group , and order 24 for the tetrahedral group ) it is hence not surprising that icosahedral symmetry plays a distinguished role . if only the symmetry axes of the icosahedral group ( 12 5-fold axes , 20 3-fold axes and 30 2-fold axes ) could be used to encode the locations of protein subunits
, only small viral capsids would occur .
viral capsids therefore also follow local symmetry axes .
this has already been observed by caspar and klug @xcite , and they classify the structure of particles that follow a particular type of local symmetry , which runs under the terminology _ quasi - equivalence _ in their work .
however , experimental evidence has shown that viruses follow also local symmetries different from these @xcite .
inspired by this , a mathematical formalism is developed here which can be used to determine the surface structures of all icosahedral particles with local symmetries .
the approach adopted here is based on the fact that the rotational symmetries of the icosahedral group are generated by the reflections that form the non - crystallographic coxeter group @xmath0 .
the locations of the corresponding planes of reflection are encoded by a set of vectors , called root vectors , that are the ( normalized ) vectors perpendicular to the reflection planes . in order to obtain further planes of reflection that correspond to local rotational symmetries and are compatible with the overall group structure ( global symmetries ) , the set of root vectors has to be extended in an appropriate way . in order to achieve this ,
affine extensions of non - crystallographic coxeter groups are used @xcite as discussed in section [ two ] .
they lead to an augmentation of the set of roots by a further root , and @xmath1-linear combinations of this enlarged set of roots define nested point sets that are subsets of generalised lattices .
the points in these sets correspond to the normal vectors of additional reflection planes and hence encode the locations of additional ( local ) symmetry axes that are compatible with the overall group structure .
the structures of icosahedral particles with local symmetries can thus be reconstructed based on an affine extension of the non - crystallographic coxeter group @xmath0 . in section [ two ] a simple tool - kit is developed based on this formalism which can be used to compute the structure of these particles .
it is easy to implement and can be used also by people who are not familiar with group theory .
applications are considered in section [ three ] .
it is important to note that the formalism presented here can be used to construct _ all _ icosahedral particles with local symmetry axes .
the reason for this lies in its connection with the projection formalism known from the study of quasicrystals @xcite and penrose tilings @xcite .
in particular , it is known that point sets with local symmetries can be obtained via a projection from a higher dimensional regular lattice such as , for example , the root lattices corresponding to indecomposable finite root systems @xcite .
all local symmetries of particles with overall icosahedral symmetry can hence be obtained via projection from the root lattice of @xmath2 @xcite .
since the root system of @xmath0 can be obtained from the root system of @xmath2 by projection @xcite , and the root lattice of @xmath2 is given by @xmath1-linear combinations of the root vectors , which after projection lead to @xmath1-linear combination of the root vectors of @xmath0 , one can equally work with @xmath1-linear combinations of the root vectors of @xmath0 themselves provided that one has a guide as to which linear combinations are important .
this information is provided by the affine extension of the group , and leads to the tool - kit for the construction of particles with local symmetries presented in subsection [ twotwo ] .
therefore , even though one could in principle construct the surface structures of the particles via projections from the root lattice of @xmath2 , a different approach is adopted here , which is less cumbersome because remnants of the higher dimensional space are only implicitly contained in the theory .
all practical computations reduce to very simple operations in three dimensions that can easily be implemented on a computer .
applications are discussed in section [ three ] .
due to the importance of icosahedral symmetry in virology , we present the theory with focus on this case . however , the theory can be used also for the derivation of the surface structures of particles with octahedral and tetrahedral symmetry , if the root systems of the corresponding reflection groups are used instead of the root system of @xmath0 .
the coxeter group @xmath0 is the only non - crystallographic coxeter group in three dimensions .
it is a subgroup of @xmath3 , the only non - crystallographic coxeter group in four dimensions , and contains @xmath4 as a subgroup @xcite .
the root systems of @xmath3 and @xmath4 can be obtained via projection from the root systems of the crystallographic groups @xmath5 and @xmath6 , respectively . as in the case of @xmath0 and @xmath2 ,
the higher dimensional crystallographic root systems in these pairings are distinguished by the fact that all root vectors are of the same length .
the elements of @xmath0 are reflections @xmath7 at planes perpendicular to the root vectors ( or roots in short ) @xmath8 .
the collection of all root vectors is called the root system , denoted as @xmath9 , and encodes all elements in the group as follows : to each @xmath10 there corresponds a reflection @xmath11 , which is a reflection at the plane perpendicular to the root @xmath8 .
the root system of @xmath0 is given by the following 30 vectors @xcite
: @xmath12 where @xmath13 and its algebraic conjugate @xmath14 are the two solutions of the quadratic equation @xmath15 . these vectors point from the centre of an icosidodecahedron to its vertices , as shown in fig .
[ sp ] on the left , and encode each the location of a plane of reflection , as shown on the right .
( left ) and the root polytope that encodes the locations of the planes of reflection ( right).,title="fig:",width=143 ] ( left ) and the root polytope that encodes the locations of the planes of reflection ( right).,title="fig:",width=143 ] the intersections of the planes of reflection mark the locations of the axes of rotational symmetry of the icosahedral group . for example , the intersection of the two planes in the figure corresponds to two axes of 5-fold symmetry , which intersect the sphere at two of the 12 5-fold vertices of the ( spherical ) icosahedron .
the other reflection planes are not shown , but their intersections with the surface of the sphere are indicated as geodesics ( spherical arcs obtained as intersections of planes through the origin with the surface of the sphere ) .
the intersections of the geodesics mark the locations of all symmetry axes , and one can hence reconstruct the locations of all 12 five - fold , 30 two - fold and 20 three - fold axes from them . in order to extend the root system in a way compatible with overall icosahedral symmetry and
hence find new planes of reflection as well as the associated axes of local symmetry , one needs to observe that the information encoded by the root system @xmath9 can further be condensed into three so - called simple roots .
they form a basis from which all other root vectors can be constructed by positive or negative @xmath16$]-linear combinations , where @xmath16=\lbrace a+\t b\,\vert\ , a , b\in\z\rbrace$ ] corresponds to an extended ring of integers .
in particular , a possible choice of simple roots in the orthonormal basis is @xmath17 the relations between the simple roots are encoded in the cartan matrix @xmath18 , @xmath19 via an affine extension this matrix is extended by an additional row and column , which encode a further root , that corresponds to an affine reflection .
it has been shown in @xcite , that the affine extended cartan matrix is given by @xmath20 and that the affine reflection act as a translation @xmath21 by the highest root @xmath22 .
the three other reflections are cyclic operations of order two and the products of any two of them correspond to rotations around the origin : @xmath23 the affine extended group is hence generated by the three reflections @xmath24 , @xmath25 , @xmath26 as well as @xmath21 .
three - dimensional points sets with local symmetries can be constructed via an iterated action of the generators of the extended group on the origin .
if the action of the translation operator @xmath21 is not restricted , space is densely filled in this way . however ,
if @xmath21 acts only a finite number of times , n say , while the action of all other operations is not restricted , point sets @xmath27 are obtained which are finite subsets of cut - and - project quasicrystals with centrally symmetric acceptance windows . for increasing @xmath28 , the patches become larger and more dense , and
correspond to cut - and - project quasicrystals with increasingly larger acceptance windows .
examples have been worked out explicitly for the two - dimensional subgroup @xmath4 in @xcite , and for @xmath0 in @xcite , where applications to carbon cage structures have been considered . due to the fact that that @xmath21 acts as a translation by the highest root , a simple expression for these point sets can be given in terms of the root system @xmath9 ( see @xcite for the case of @xmath0 ) : @xmath29 i.e. @xmath27 consists of all points that are obtained as an @xmath30-linear combination of up to @xmath28 roots in @xmath9 . these sets form the starting point for the construction of icosahedral particles with local symmetries in the next subsection .
the sets @xmath27 in ( [ sn ] ) decompose into disjoint subsets @xmath31 of vectors of equal length that are invariant under the action of @xmath0 : @xmath32 where @xmath33 denotes the number of different subsets of vectors of equal length in @xmath27 .
these subsets are of different cardinalities , and a given set @xmath31 may increase in cardinality for increasing @xmath28 , i.e. one has the set inclusion @xmath34 for @xmath35 .
@xmath31 encodes the surface structure of a particle that has local symmetry axes compatible with the overall group structure . since only the orientations but
not the length of the vectors in the set is important for the reconstruction of the locations of the local reflection planes , two sets @xmath31 correspond to the same particle if they contain the same vectors modulo length .
therefore the range of different ( modulo length ) sets @xmath31 corresponds to the spectrum of different icosahedral particles with local symmetries .
the tool - kit given here provides a method for the systematic evaluation of this information and hence for the construction of the surface structures of all such particles . by definition
, @xmath36 corresponds to the root system @xmath9 .
the reflection planes orthogonal to the roots intersect precisely at the symmetry axes corresponding to the ( global ) symmetries of the icosahedral group , and it is hence a structure without local symmetries . therefore , local symmetries are encoded in the point sets @xmath27 with @xmath37 . for each @xmath28 ,
the sets @xmath31 in ( [ orbit ] ) need to be evaluated individually .
this is done by an inductive process starting with the case @xmath38 .
it is presented in the form of a tool - kit which facilitates easy computational implementation and can be used also by people not familiar with the group theoretical set - up : 1 .
add any two vectors in ( [ icosh3 ] ) and group these vectors into sets @xmath39 , @xmath40 , of vectors of equal length .
for each @xmath41 , compute the planes through the origin that are perpendicular to the vectors in @xmath39 according to the equation @xmath42 for any vector @xmath43 with cartesian components @xmath44 .
the intersections of these planes determine the locations and types of the corresponding local symmetry axes .
they may either be computed analytically or graphically displayed with a visualisation tool such as maple .
2 . compute the structure of @xmath27 inductively based on the structure of @xmath45 as follows : add any n vectors in ( [ icosh3 ] ) and group them by their length into sets @xmath46 . if the length of the vectors in a set @xmath46 corresponds to that of one of a sets at iteration step @xmath47 ,
@xmath45 say , form a new set that contains both sets , i.e. introduce a new set @xmath48 . otherwise , add the set as a new set to the list , i.e. @xmath49 , where the index @xmath50 is chosen such that sets are numbered consecutively .
the collection of sets obtained in this way corresponds to the sets @xmath31 , @xmath51 in ( [ orbit ] ) . for each of them , compute the corresponding symmetry planes according to ( [ symplane ] ) as before .
applications of this tool - kit are discussed in section [ three ] .
note that the intersections of the reflection planes with the surface of the sphere lead to spherical tessellations of the surface of the sphere .
the set inclusions of the type @xmath52 for @xmath53 imply that the spherical tiling ( tessellations ) corresponding to these sets are related by _
selfsimilarity _ : larger copies of the tiles ( areas ) implied by the spherical tiling related to @xmath54 are replaced by an integer number of rescaled versions thereof in the spherical tiling related to @xmath55 .
an example is shown in fig .
[ triangles ] , where each of the four triangles in the original triangulation ( with vertices shown as blue dots ) has been replaced by four rescaled ( smaller ) versions in the refined triangulation .
the tiling is shown only in one of the 20 faces of the spherical icosahedron ( or , equivalently , in three copies of the fundamental domain of the icosahedral group ) , because the remainder of the tiling follows from this via the actions of the icosahedral group .
planes represented by lines connecting the blue dots arise from the vectors in the set @xmath54 , while planes represented by dashed lines connecting the red dots correspond to the vectors that are contained in @xmath55 in addition .
are shown in red.,width=188 ] this phenomenon is also known from the study of quasicrystals @xcite . since quasicrystals are obtained from a projection formalism that is related to this construction as discussed in the introduction , this is not surprising .
in this section it is demonstrated how the theory can be applied in virology . in subsection [ threeone ]
the tool - kit is applied to derive the surface structures of new viral particles with local symmetries . in subsection [ threetwo ]
it is shown that the theory answers , as a by - product , an open question in virology : it explains why only certain types of capsomeres ( protein clusters ) are observed in viral capsids .
the local symmetries of particles with icosahedral symmetry can be obtained from the root system of the non - crystallographic coxeter group @xmath0 via the tool - kit presented in the previous section . according to this ,
we start by considering the set @xmath56 of all vectors obtained via a linear combinations of any two root vectors in ( [ icosh3 ] ) and decompose this set into subsets @xmath39 of roots of equal length .
the types of the local symmetry axes that are obtained as intersections of the reflection planes encoded by the vectors in a set @xmath39 depend on the relative orientations of these vectors , as can be seen from the examples discussed below . 1 . _ a particle with local 3-fold axes : _ + consider the subset @xmath39 of @xmath56 given by linear combinations of any two vectors pointing to vertices of the icosidodecahedron with a joint edge , such as , for example , the vertices shown in blue in fig .
[ ck ] . the resulting vectors intersect the edges at the points shown in red . in order to reconstruct the locations of the reflection planes orthogonal to these vectors , note
that the organisation of the edges reflects the subgroup structure of @xmath0 : they are grouped in decagonal arrangements according to copies of the subgroup @xmath4 .
each such decagon encodes 10 root vectors of @xmath0 , which in turn encode the symmetry planes that intersect at a global 5-fold vertex and its antipode on the sphere .
therefore , the bisection of the edges in the root vector picture in fig
. [ ck ] on the left correspond to symmetry planes that bisect the angles between two planes meeting at a 5-fold vertex at an angle @xmath57 .
the geodesics corresponding to the intersection of such planes with the sphere are shown in fig .
[ ck ] on the right superimposed on the spherical icosahedron . in order to keep the presentation transparent , only the geodesics in one of the 20 faces of the spherical icosahedron
are shown : geodesics arising from bisections of the angles within this spherical triangle are shown as dashed lines ; moreover , the geodesics arising from bisections of angles in different spherical triangles ( faces ) are shown as dotted lines in order to show their impact on the spherical triangle under consideration .
+ + note that the intersections of these geodesics determine the locations of the local symmetry axes : there are three local symmetry axes of order six in the spherical triangle .
since this spherical triangle contains three copies of the fundamental domain of the icosahedral group , this setting corresponds to one new local 6-fold symmetry axis that is reproduced 60 times under the action of the icosahedral group .
note that the planes of symmetry , respectively the geodesics corresponding to them , not only fix the locations of the symmetry axes but also the relative orientation of the planes of symmetry and hence the surface structure of the corresponding viral capsids is completely determined by this formalism .
2 . _ a particle with local 5-fold axes : _ + in order to obtain a particle with local 5-fold symmetry axes
, a different subset @xmath39 of @xmath56 has to be considered , which inherits its order of symmetry from the global 5-fold axes . for this , consider linear combinations of root vectors that are located on two different @xmath4-orbits , as for example the vectors pointing to the vertices shown in blue in fig .
[ rt1 ] on the left .
these vertices are located on two @xmath4-orbits that are shown in green and red , respectively , and intersect at the vertex shown in yellow .
[ rt1 ] on the right shows how this picture translates into reflection planes .
in particular , part of the geodesics corresponding to the reflection planes perpendicular to the 10 root vectors pointing to vertices on the green , respectively red , @xmath4-orbit are shown schematically in green , respectively red , around 5-fold vertices in the spherical tiling .
the geodesics corresponding to the two root vectors pointing to vertices marked by blue dots in the figure on the left are shown in blue in the spherical tiling on the right .
[ rt1 ] hence provides a dictionary on how to translate the geometry of the root vectors into geodesics on the sphere .
+ -orbit structure in the root system ( left ) into geodesics corresponding to reflection planes ( right).,title="fig:",width=188]-orbit structure in the root system ( left ) into geodesics corresponding to reflection planes ( right).,title="fig:",width=188 ] + based on this , it is possible to construct the axes of local five - fold symmetry as follows .
the linear combinations of any two root vectors corresponding to the blue dots in fig .
[ rt2 ] ( left ) corresponds to a vector that points from the centre of the icosidodecahedron to any of the vertices marked in red in fig .
[ rt2 ] ( left ) . reconstructing the corresponding planes of symmetry as before leads to a network of geodesics with intersections of order 10 rather than 5 .
however , in order to obtain the spherical tiling that represents the surface structure of the viral particle , note that also the global 5-fold axes appear as 10-fold axes in this representation , because the angles between the five reflection planes at each 5-fold vertex are bisected by the reflection planes originating at the 5-fold vertex at its antipode on the sphere .
similarly , the local symmetry axes have to be interpreted as 5-fold symmetry axes in disguise , and every second geodesic has to be ignored .
this leads to two different spherical tilings , shown in fig .
[ rt2 ] in the middle and on the right , respectively .
+ + in order to visualise the structure of the viral capsids encoded by these tilings , the rules of the tiling approach need to be used @xcite : protein subunits are located in the corners of the tiles that meet at the local symmetry axes and are represented schematically as dots .
the colours of the dots encode which proteins are mapped onto each other under the action of the overall symmetry group ; the proteins themselves are identical and hence indistinguishable
. tiles with two and three dots represent , respectively , dimer- and trimer - interactions , i.e. interactions between , respectively , two and three protein subunits .
they are visualised schematically as spiral arms and may correspond for example to c - terminal arm extensions of the proteins . with this interpretation , the tiling in fig .
[ rt2 ] on the right leads to the surface structure shown in fig .
[ surfaces ] .
+ ( right).,width=181 ] + note that the predictions of this tiling differ both by the orientations of the protein clusters ( pentamers ) , which are rotated by an angle of @xmath58 with respect to the other alternative , and by the types and locations of the intersubunit bonds that stabilize the capsid .
the tiling in the middle of fig .
[ rt2 ] corresponds to the tiling in @xcite and represents viral particles in the family of papovaviridae .
viral particles following the alternative in fig . [ surfaces ]
are not known at present .
note that these sets are only two examples of sets in the decomposition of @xmath56 .
they have been chosen because they are the most instructive examples
. there exists , for example , also a subset that does not introduce any local symmetry axes : it corresponds to linear combinations of each root vector with itself .
these vectors point to the vertices of a larger copy of the icosidodecahedron , and the set hence encodes only the global symmetry axes ( like the root system itself ) and hence contains no new information .
while the geometry of the particle in item 1 is known already from the caspar - klug classification ( it belongs to a @xmath59 capsid ) , the geometries in item 2 are new .
they correspond to particles with only 5-fold axes of symmetry ( global and local ) . due to overall icosahedral symmetry
, any local 5-fold axis occurs in multiples of 60 , so that they are indeed the smallest geometries corresponding to particles with local five - fold axes .
a classification of all icosahedral particles with local symmetries based on the theory developed here is in progress @xcite .
the centres of the capsomeres in viral capsids correspond to the local and global axes of symmetry .
therefore , the spectrum of different capsomeres that may occur in nature is determined completely by the orders of the symmetry axes that can occur from a mathematical point of view .
the group theoretical formalism developed here implies that the orders of all local symmetry axes must correspond to multiples of those of the global ones , because they inherit their symmetry from the structure of the root system due to the fact that they arise from @xmath1-linear combinations of the root vectors .
therefore , the orders of the local symmetry axes must be @xmath1-multiples of 2 , 3 or 5 , and local symmetry axes of , for example , order 7 can not occur .
moreover , since smaller particles are obtained from linear combinations of only a few root vectors , multiples of more than twice the order of the global axes do not occur for them .
hence , for all smaller icosahedrally symmetric particles , which form the majority of viruses observed to date , one would expect a spectrum of local symmetries of order 3 , 5 , 6 and 10 .
all of these are known to occur : order three has been observed for bacteriophage ms2 @xcite , order five for polyomavirus @xcite and simian virus 40 @xcite , order six in all caspar - klug cases , and order 10 for l - a virus of yeast @xcite .
all these cases follow the construction principle outlined above .
the formalism presented here is used in @xcite to describe the surface structures of viral capsids in terms of tilings .
the vertex sets of the tilings contain the points , that correspond to the local and local symmetry axes , and mark the locations of the capsomeres .
the edges of the tiles follow the geodesics that arise as intersections of the planes of ( local and global ) symmetry with the surface of the sphere . in this sense ,
the tiling picture and the group theoretical picture in terms of local symmetry axes are equivalent ways of modelling the surface structures of viral capsids .
this may suggest that the tiles are of secondary importance and that the symmetry axes should be the focus of attention .
however , in this way , important additional information on the structure of viral capsids would be neglected as tiles represent the interactions between subunits belonging to different capsomeres and are hence biologically meaningful objects . in particular , the measures and shapes of the tiles are correlated with the strengths and types of the intersubunit bonds they represent @xcite .
their structure is hence very important information in order to determine the types of the interactions ( i.e. whether they are trimer- or dimer - interactions , that is interactions between three , respectively two , protein subunits ) , as well as their exact locations .
this information is crucial input for the construction of assembly models @xcite , that investigate how viral capsids are assembled from their building blocks , as well as for the modelling and classification of additional structures such as the covalent bonds responsible for crosslinking @xcite .
financial support via an epsrc advanced research fellowship is gratefully acknowledged .
caston , i.r , trus , b.l , booy , f.b , wickner , r.b , wall j.s @xmath60 steven , a.c .
( 1997 ) structure of l - a virus : a specialized compartment for the transcription and replication of double - stranded rna . _ the journal of cell biol . _ 138 , 975986 .
moody , r. v. and patera , j ( 1993 ) , the @xmath6 family of quasicrystals , in : _ proc .
nato asi noncompact lie groups and some of their applications _ ,
san antonio , eds .
tanner , e.a . and
wilson , f. , nato asi series c , vol.429 , pp.341 . | a formalism is developed which allows to determine the locations of all local symmetry axes of three - dimensional particles with overall icosahedral symmetry .
it relies on the fact that the root system of the non - crystallographic coxeter group @xmath0 encodes the locations of the planes of reflection that generate the discrete rotational symmetries of the particles . via an appropriate extension of the root system , new planes of reflection
are introduced which determine local axes of rotational symmetry .
an easy - to - implement formalism is derived that allows to compute the surface structure of any three - dimensional icosahedral particle with local symmetries .
it can be used also for particles with overall octahedral and tetrahedral symmetry in conjunction with the root systems of the corresponding reflection groups .
applications to viruses are discussed explicitly .
it is shown that the concept of quasi - equivalence in caspar - klug theory corresponds to the special case of local six - fold symmetry axes contained in the theory developed here , and the corresponding geometries can hence be obtained with this formalism based on the root system of @xmath0 . moreover , as a by - product , the theory answers the long - standing open question why only certain types of capsomeres , i.e. clusters of protein subunits , are observed in the surface structures of viruses .
since the types of the capsomeres are determined by the orders of the local symmetry axes on which they are located , the possible types of capsomeres are restricted by the spectrum of local symmetry axes allowed by the theory .
based on this we determine the spectrum of all capsomere types that may occur in viral capsids and give explicit examples for the lower - order cases .
+ + + |
a progress in telecommunications , using of new methods in data transfer , creation of medical equipment are in need of the development of microelectronic devices for generation , commutation and detection of different kinds of signals .
the most actual problem is an engineering of equipments working in terahertz and far infrared range of electromagnetic waves ( see , for example , @xcite ) .
one way to determine the characteristics of high frequency electromagnetic waves ( amplitude , frequency , phase ) is a study of the direct current response which can occur under certain conditions in solid - state structures with non - linear properties ( for example , ) . among the similar phenomena
it should be noted a circular photogalvanic effect ( cpge ) ( see reviews @xcite ) , which was investigated in surface layers of bulk materials @xcite , in thin films @xcite , quantum wells @xcite , in graphene @xcite . without going into details , this effect is consisted in the appearance of a direct current under the influence of elliptically polarized electromagnetic wave with oblique incidence to the surface of the sample .
the current flows perpendicularly to the projection of wave vector on the plane of the sample . by the magnitude of direct current
one can draw a conclusion about properties of radiation as well as properties of a considered structure @xcite .
effect is directly associated with the presence of spin of charge carriers and it was mostly explored in materials with hamiltonian containing the term which is linear by a quasi - momentum ( so - called rashba @xcite or dresselhaus @xcite terms ) .
such hamiltonians take in consideration a spin - orbital interaction immediately @xcite .
most of the works on cpge are devoted to research in visible and near infrared range of electromagnetic waves , where inter - band @xcite or inter - subband @xcite optical transition are essential for kinetic properties of considered structures . in range of low frequencies of radiation , where only indirect intraband transitions without spin - flip are possible
, the effect is not appear @xcite .
it should be noted that even though photon energy is not enough for realization of direct transitions , there are possible transition processes through virtual intermediate state @xcite , which is placed in another band , and cpge is feasible , but these transitions are second - order processes and the direct current magnitude in this case is noticeably less .
cpge is related to the adsorption of circularly polarized electromagnetic wave which is resulting in optical orientation of spin because of angular momentum transfer from photons to electrons @xcite .
a preferential direction appears in initially homogeneous electron momentum distribution which leads to a direct current arising . at oblique incidence of wave
this preferential direction is defined by the projection of wave vector on the plane of the sample . at normal incidence of elliptically
polarized wave a direct current is usually associated not so with symmetry breaking in distribution of charge carriers momenta as with the inhomogeneity of the sample @xcite .
a preferential direction in material can be made by another way so , for example , applying a constant external electric field .
in this situation a direct current generation is possible in direction perpendicular to the constant electric field under the influence of elliptically polarized electromagnetic wave , which is incident normally to the surface of the sample , when the frequency of wave is small compared with the band gap width .
herewith a main contribution to direct current component should be made by indirect transitions between the states within one band , that are the first - order processes , and the effect should not depend on spin . in papers @xcite
the effect of a direct current appearance in graphene and superlattice on the base of graphene in perpendicular to a constant field direction is investigated , when the elliptically polarized electromagnetic wave is incident normally to the surface of the sample ( the problem geometry is shown in [ fig:1 ] ) .
the effect arising relates to a material spectrum non - additivity @xcite . in @xcite on the base of monte carlo modelling it was shown that the inelastic scattering of charge carriers on optical phonons makes a main contribution in this effect . in present paper
it is made an attempt to investigate on the base of semiclassical monte carlo simulations the effect of generation of a direct transverse current under a simultaneous influence of elliptically polarized wave and constant electric field in material described by rashba hamiltonian .
also we consider this problem in the approximation of a constant collision frequency and define the possibility of application of this approach .
let us to examine a sample which surface is irradiated by normally incident elliptically polarized electromagnetic wave with components of @xmath0 and @xmath1 .
a constant electric field @xmath2 is applied along the surface of the sample ( fig .
1 ) . the result electric field vector takes a form : @xmath3 we shall consider a two - dimensional material , described by the rashba hamiltonian : @xmath4 where @xmath5 are the pauli matrices : @xmath6 energy spectrum can be presented of : @xmath7 here @xmath8 is a rashba parameter @xcite , @xmath9 is a quasi - momentum vector of electron . the expression ( [ eq : eq2 ] )
describes two subbands , in which a conductivity band splits after removal of the spin degeneracy @xcite . in present work
we shall investigate transitions within one subband only , so we , to be more specific , shall choose in ( [ eq : eq2 ] ) plus sign .
successive accounting of the influence of interband or intersubband transitions on the effect will be a subject of a separate study .
let us to find a direct component of the current density along the @xmath10 axis : @xmath11 where @xmath12 is a velocity projection , @xmath13 is a surface concentration of charge carriers ( @xmath14 , @xmath15 - surface area of crystal , @xmath16 - number of charge carriers ) , the angle brackets are denoted an averaging on ensemble and time .
strictly speaking , for carrying out averaging one has to know an electron distribution function , which requires the solution of kinetic equation .
this approach we use in sect .
3 where boltzmann kinetic equation in a constant collision frequency approximation is solved . because of mathematical complexity of problem the expression of direct current density
is obtained in assumption of relative smallness of linear on quasi - momentum term in expression of energy spectrum in compare with quadratic term .
this solution gives an opportunity to qualitative explain some features of the direct current density dependence on components of the electric field .
but the approach of constant relaxation time does nt take in account the microscopic processes of charge carriers scattering , so for definition of relative contribution of different types of scatterers on the direct current density we use monte carlo modelling .
method monte carlo allows us to perform an averaging of the velocity without the knowledge of an explicit form of the distribution function .
note , that used below monte carlo method is essentially a method of solving of boltzmann kinetic equation ( mathematically rigorous justification of this approach is presented , for example , in @xcite ) , that is why all restrictions , related to the application possibility of the kinetic equation ( transition probability assumed to be independent on applied fields , predominantly accounting pair collisions , etc . ) concern to used approach too .
the use of the monte carlo simulations for solving of physical kinetics problems ( see , for example , @xcite ) is a productive technique . using a fairly simple and easy to parallelized code
, one can get a sufficiently accurate and reasonable physical results .
[ sec:3 ] is devoted to a description of the modelling method .
[ sec:4 ] contains a calculation of charge carriers scattering probability on acoustical and optical phonons in material with rashba hamiltonian .
[ sec:5 ] is devoted to a description of the modelling of charge carriers distribution on the quasi - momentum absolute value and polar angle in two - dimensional quasi - momentum space , which is used in monte carlo simulations . in sect .
[ sec:6 ] the results of monte carlo simulations are presented .
also in this section there is a comparison of monte - carlo simulation results and results of the investigation of the problem on the base of boltzmann kinetic equation in approximation in constant collision frequency .
we shall investigate boltzmann kinetic equation in approximation of a constant collision frequency : @xmath17 here @xmath18 is an equilibrium distribution function , @xmath19 is a mean collision frequency .
the solution of ( [ eq : eq4 ] ) is a non - equilibrium distribution function .
it can be found by a method of characteristics and takes a form @xmath20 here @xmath21 is a solution of semi - classical equations of motion @xmath22 the solution of ( [ eq : eq6 ] ) is @xmath23 where we introduce denotations @xmath24 we shall investigate non - degenerated electronic gas with equilibrium distribution function @xmath25 here @xmath26 is absolute temperature , @xmath27 is a normalization constant , @xmath28 @xmath29 is an error function , @xmath30 is a dimensionless parameter responsible for the spin - orbital interaction .
so the current density along @xmath10 axis can be found by expression @xmath31 let us to carry out a substitution @xmath32 , @xmath33 .
so the expression ( [ eq : eq11 ] ) is rearranged in the form @xmath34 in dimensionless denotations @xmath35 the current density is @xmath36 here @xmath37 at a temperature @xmath38 , assuming @xmath39 , @xmath40 , the dimensionless rashba parameter @xmath41 , so in a first non - vanishing approximation on @xmath42 we can keep only a linear term in the expression of the direct current density . after carrying out the taylor expansion on @xmath42 the expression under integral , passing into polar system of coordinates , remembering a subsequent averaging of the current density on time we give next expression for the current density : @xmath43 @xmath44 according to cauchy inequality @xmath45 , so using @xcite ( p.197 , 1.5.20.8 , 1.5.20.10 ) we receive : @xmath46.\end{aligned}\ ] ] here @xmath47 are complete elliptic integrals of the first and second kind , respectively , and @xmath48 is an elliptic integral of the third kind .
after substitution ( [ eq : eq17 ] ) into ( [ eq : eq15 ] ) we get the expression of direct current density along @xmath10 axis which can be investigated further numerically : @xmath49\right\rangle}_t .
\end{aligned}\ ] ] angle brackets in a formula above denotes an averaging on a large in compare with a period of the elliptically polarized wave time interval . because of time periodicity of the considered electric field we can perform averaging only by a period of wave .
so for examine the dependence of the direct current component along @xmath10 axis we should numerically take three integrals ( on absolute value of quasi - momentum @xmath50 , time @xmath51 and @xmath52 ) . on a dimensionless component of electric field strength of elliptically polarized wave @xmath53 .
red curve correspond to @xmath54 , blue curve to @xmath55 , black curve to @xmath56 . ] on a dimensionless component of electric field strength of elliptically polarized wave @xmath57 .
red curve correspond to @xmath58 , blue curve to @xmath59 , black curve to @xmath60 .
] on the phase shift @xmath61 between the components of elliptically polarized wave .
red curve correspond to @xmath62 , blue curve to @xmath63 , black curve to @xmath64 . ] figures [ fig:2 ] and [ fig:3 ] illustrate the dependence of the transverse direct current density @xmath65 on components of the electric field strength of elliptically polarized wave @xmath53 and @xmath57 , respectively . value @xmath66
( @xmath67 ) corresponds to approximately 10 sgse unit at a temperature @xmath68 , cyclic frequency of incident wave @xmath69 .
rashba parameter is assumed to be @xmath70 . in section [ sec:6 ] on the base of monte carlo simulation we find a mean collision frequency @xmath71 , so graphics on figures [ fig:2 ] and [ fig:3 ]
are plotted in assumptions @xmath72 .
also , to be more specific , we assume @xmath73 .
the dependencies of the direct current on a constant component of electric field @xmath74 are similar ones presented on figures [ fig:2 ] and [ fig:3 ] , and they are shown in section [ sec:6 ] , where the results of the constant collision frequency approach and the monte carlo modelling are compared . in a low - amplitude area of electric field
the direct current density is linear with respect to each component @xmath53 , @xmath57 , @xmath74 .
this result is in according with the same obtained previously in similar problem geometry in graphene analytically @xcite and by monte carlo simulation @xcite .
the effect is cubic by magnitudes of electric field strengths which is closely related to mutual influence of charge carriers motions in transverse each other directions in material with non - additive energy spectrum .
a non - monotonic character of a direct current density dependence on the components of the electric field can be explained that the effect is result of a mutual interaction of three motions ( under the influence of two alternative and one constant forces ) .
so if one of the forces became much more than two another , the effect decreased .
another cause of decreasing of the direct current with rising of one component of the electric field is that in the states where an electron has small values of quasi - momentum , a term in the energy spectrum , linear with respect to a quasi - momentum , prevails over the quadratic term , and a non - additivity of the energy spectrum is significant . with rising of a mean value of a quasi - momentum a relative contribution of the linear term
is reduced , an influence of the energy spectrum non - additivity decreases and the direct component of a transverse current also diminishes .
the dependencies of the direct current density on a phase shift between the components of wave is very close to cosine ( fig .
[ fig:4 ] ) . with growth of a collisional frequency
the direct current component monotonically decreases .
consequently , on the base of examination of boltzmann kinetic equation with the collision term in a form of constant relaxation frequency we calculate a direct current density ( in perpendicular direction to a constant component of electric field strength ) in material with rashba hamiltonian under the influence of elliptically polarized electromagnetic wave , which is incident normally to the surface of the sample .
this approach allows us to get the dependence of direct current on components of electric field , collision frequency and phase shift between the components of elliptically polarized wave .
it is established that essential contribution in effect appearance belongs to energy spectrum non - additivity . but
the method of constant relaxation time ca nt immediately take in account microscopic mechanisms of charge carriers collisions on the lattice heterogeneity .
meanwhile , only a non - additivity ca nt lead to appearance a preferential direction in a sample .
direct current should arise as result of momentum and energy interchange between electron gas and lattice under the influence of electric field .
next sections are devoted to clarification of relative influence of different types of scattering on appearance of direct current on the base of semi - classical monte carlo simulations .
let us to watch over an electron which moves in electric field and is capable to be scattered by lattice irregularities .
we assume that during the intervals between collisions the particle motion obeys the laws of classical mechanics and is described by the equations of motion @xmath75 the time intervals between collisions and the final momentum acquired by the particle after the collision , we model by the random variable , the distribution of which will be determined by the probabilities of transition from one state to another as a result of scattering .
the transition probabilities are calculated from quantum mechanical considerations .
calculating the momentum component for single electron at each point of time ( taking into consideration the motion in electric field and scattering ) , we calculate the electron velocity in each point of time . repeating many times similar calculations for different initial conditions ( one might say , for different particles ) , and performing an averaging across the set of initial conditions and with respect to a time
, we obtain an expression for the mean velocity , and then , using ( [ eq : eq3 ] ) , we calculate a direct current density .
as it was shown by simulation @xcite and as it follows from physical reasons , an inelastic scattering should play a key role in causing of transverse rectification effects .
an energy exchange between the lattice and electron gas leads to demonstration of the mutual dependence of charge carriers movements in directions perpendicular to each other , potentially inherent in the materials with non - additive energy spectrum .
that is why during the farther consideration of effect of direct current appearance in material with rashba hamiltonian under the conditions of elliptically polarized electromagnetic wave we shall take in account two mechanisms : scattering of charge carriers on optical phonons ( substantially inelastic ) and scattering on acoustical phonons ( which is close to elastic scattering ) .
we shall need the expressions of total transition probability @xmath76 from the state with momentum @xmath77 to the state with any another allowable momentum value .
the sect .
[ sec:6 ] is devoted to calculation of these probabilities .
this section describes the considerations , on the base of which the time of electron scattering on phonons is modelled and the scattering mechanism is chosen . a mean free time can be found by the next considerations set forth , for example , in @xcite .
suppose , first , that the probability of an electron scattering in a time interval is proportional to the interval , and , second , the probability of collisions per unit time is independent of time .
let @xmath78 is a probability that a particle moving without collisions over a period of time @xmath79 .
consider the event @xmath80 , consisting in the fact that electrons move without scattering in the interval @xmath81 .
this event can be considered as a combination of two events : the event @xmath82 , which consists in the fact that the electron moves without scattering over a period of time @xmath83 , and the event @xmath84 consisting in the fact that the electron is moving without being scattered in time interval @xmath85 . since we assume that the probability of the free motion of a particle depends on the duration of the period of time , but not on the initial ( or final ) point of time , the events @xmath82 and @xmath84 are independent , so @xmath86 on the other hand , @xmath87 the probability of free movement for the time @xmath88 can be expressed by the scattering probability @xmath89 for the same time : @xmath90 for @xmath91 one obtains next equation : @xmath92 the solution of this equation is a function @xmath93 where the constant @xmath82 is determined by the normalization condition @xmath94 thus , the probability that an electron will move without scattering within the time @xmath83 @xmath95 there @xmath89 is a total probability of charge carriers scattering per unit time .
this probability depends primarily on the electron quasi - momentum : @xmath96 .
the quantity @xmath97 in the exponent , in fact , is the probability of charge carriers scattering in the time interval duration @xmath51 .
when the sample is placed in an electric field , an electron moves between collisions according to the equation of motion ( [ eq : eq23 ] ) , so that the momentum ( and the probability of scattering ) begins to depend on the time .
generalization of ( [ eq : eq30 ] ) in this case is to replace @xmath98 ( @xmath99 is a vector - potential ) , @xmath100 so the probability that an electron will not experience a collision during a period @xmath51 when it moves under the influence of the external field is @xmath101 expression ( [ eq : eq33 ] ) can be used for drawing the time between successive scatterings .
we assign to @xmath91 a value @xmath102 uniformly distributed in the interval @xmath103 .
then the time @xmath51 of motion of an electron under the influence of fields between two successive collisions can be determined by solving the equation @xmath104 solving of ( [ eq : eq34 ] ) is a complex process . therefore , to simplify the calculations , we use techniques developed in @xcite . we introduce a fictitious self - scattering process , which does not change the momentum of a particle @xcite .
let the probability of transition between the states @xmath77 and @xmath105 as a result of this process @xmath106 value @xmath107 in ( [ eq : eq35 ] ) can be chosen arbitrarily , since the state of the particle as a result of self - scattering does not change
. in ( [ eq : eq34 ] ) @xmath76 is the probability of scattering due to all the processes taken into consideration . then , with the introduction of self - scattering , ( [ eq : eq34 ] ) takes a form : @xmath108 considering @xmath109 where @xmath110 is some constant , we obtain the condition for the determination of the collision moment in the form of @xmath111 the value @xmath110 should be chosen in such a way that in the region of interest values of the momentum the value @xmath107 was positive .
there is another approach to the calculation of the mean free time @xcite .
as before , we assume that the probability of scattering particles in the time interval @xmath79 is proportional to the length of this interval : @xmath112 here @xmath89 is the probability of charge carriers scattering per unit time .
let the random variable @xmath102 is a time of free path .
we introduce the distribution function of this value @xmath113 the probability that the time of free path @xmath114 is less than @xmath51 : @xmath115 then the probability of the event @xmath80 , consisting in the fact that the time of free path is in between @xmath116 , will be equal @xmath117 on the other hand , the event @xmath80 can be represented as a superposition of two events : @xmath82 `` the particle is moving without scattering over a period of time @xmath118 '' and @xmath84 `` particle scatters in @xmath116 '' .
these events are independent .
an independence of event @xmath82 on the event @xmath84 is unquestioned .
an independence of event @xmath84 on @xmath82 is clear from the following .
the fact that the scattering of particles is occurred in the interval @xmath116 ( event @xmath84 ) by our assumption depends on the length of the interval @xmath88 and on the state of the particles ( primarily , on its quasi - momentum ) .
electron scattering probability per unit time @xmath89 depends on the quasi - momentum of particle , which in the case of external fields applying is a function of time , but there is no explicit dependence of the probability on the time measured from the previous collision , so there is no explicit dependence of the event @xmath84 on the initial time point of interval @xmath116 , and , hence , there is no the dependence of the event @xmath84 on the event @xmath82 .
therefore the probability of an event @xmath80 can be represented as a product of probabilities of the events @xmath82 and @xmath84 : @xmath119 obviously , the probability of event @xmath82 is a complementary quantity for @xmath113 : @xmath120 the probability of event @xmath84 is defined by ( [ eq : eq38 ] ) . combining ( [ eq : eq38 ] ) , ( [ eq : eq40 ] ) , ( [ eq : eq41 ] ) , ( [ eq : eq42 ] ) , we obtain the following relationship : @xmath121 thus , the distribution function of the time of free path satisfy the differential equation @xmath122 a solution of that equation in case of @xmath123 takes a form : @xmath124 in modelling the expression ( [ eq : eq45 ] ) can be used to determine the time of free path , assuming @xmath113 is equal to uniformly distributed on the interval @xmath125 random variable @xmath102 .
thus , the time of free path is determined by solving the equation @xmath126 the expression ( [ eq : eq46 ] ) is more preferable then the expressions ( [ eq : eq34 ] ) , ( [ eq : eq37 ] ) , since during the derivation of ( [ eq : eq34 ] ) we first assumed @xmath89 was independent on time , and then we generalized ( [ eq : eq30 ] ) ( not so strict ) for the case of an implicit dependency of the scattering probability @xmath127 on time . the expression ( [ eq : eq37 ] ) takes into account the dependence of the time of free path on the scattering probability only indirectly . for practical calculations
it is convenient to transform the expression ( [ eq : eq45 ] ) : @xmath128 left part in ( 42 ) is the probability of the fact that the time of free path is greater than @xmath51 , and this probability one can simulate by a uniformly distributed random variable @xmath129 : @xmath130 in a right part of ( 42 ) under the exponent integral is taken numerically ( trapezoidal method ) : @xmath131 where @xmath132 is a time step ( we have omitted the quote in the designation of uniformly distributed random variable ) . since a right part of ( [ eq : eq49 ] ) is a monotonous decreasing function , we find a time of free path according to @xmath133 where @xmath13 is minimal number of steps required to satisfy the condition @xmath134 taking the logarithm of both sides of ( [ eq : eq51 ] ) , one gets following condition @xmath135 which is used in this paper to perform calculations . playing a random variable @xmath102 uniformly distributed in the interval @xmath136
$ ] , we find a moment of collision using ( [ eq : eq52 ] ) .
then one should select the scattering mechanism .
thereto one more random variable @xmath137 , uniformly distributed in the interval @xmath136 $ ] , is played out . in a case when only two mechanisms of scattering are supposed ( scattering by acoustic and optical phonons ) , the choice of the scattering mechanism is based on the next obvious condition : one assumes that scattering was happened by optical phonon , if @xmath138 and one assumes that electron was scattered by acoustic vibrations of the lattice in a case if @xmath139 defining the scattering mechanism , we calculate the momentum after the collision on the base of the laws of conservation of energy and momentum , further we estimate the velocity at any time during the motion of an electron in an external field and define a new collision moment .
let us to consider a probability of the electron transition from the state with a quasi - momentum @xmath77 to the state @xmath105@xcite : @xmath140 here @xmath141 - is the operator of electron - phonon interaction ( it depends on type of phonons ) , matrix element @xmath142 is taken on wave function of a system `` electron + phonons '' , @xmath13 is a number of phonons in a chosen state , @xmath143 is an energy of the system `` electron + phonons '' in this state .
the operator of electron - phonon interaction takes a form @xcite @xmath144 where @xmath145 and @xmath146 are , respectively , birth and delete operators of phonon with a branch number @xmath137 and a quasi - momentum vector @xmath147 .
wave function @xmath148 relates to a system of non - interacting electron and phonon gases , so it can be presented in a form of the product @xmath149 here @xmath150 - is an electron wave function , @xmath151 - is a phonon wave function which is equal to a wave function of harmonic oscillator .
so the expression of matrix element can be presented as @xmath152 here @xmath153 in a case of a quadratic dispersion law , with regard to only long - wavelength lattice vibrations , one can take the electron wave function in a form of plane wave @xmath154 matrix element ( [ eq : eq58 ] ) , taken on this function , is well - known and it is presented in , for example , in @xcite . in material , which is described by rashba hamiltonian , the electron wave function is given by the next expression ( in long - wavelength approximation too ) @xcite : @xmath155 @xmath156 is an angle between the components of quasi - momentum .
we chose a top subband , so the wave function becomes the next : @xmath157 consequently , the expression of a scattering probability should differ from the expression , depicted in @xcite , only by values of integrals ( [ eq : eq59 ] ) .
after simple transformations one can obtain : @xmath158 where @xmath159 - is a matrix element , taken on plane waves ( [ eq : eq060 ] ) .
so a square of the matrix element , immediately incoming into ( [ eq : eq55 ] ) , is @xmath160 in our subsequent investigations ( during the calculation of probability of electron transition from the state @xmath77 to every possible states with quasi - momentum @xmath161 ) we shall assume @xmath10 axis directed along the vector @xmath77 , so angle difference @xmath162 should be equal to polar angle @xmath163 : @xmath164 .
thus , formulae of scattering probability in a case of calculation of matrix elements on eigenfunctions of rashba hamiltonian should differ from that , calculated on plane waves , only by coefficient @xmath165 ( taking into account , of course , that we consider the two - dimensional problem ) .
general expression for charge carriers probability takes a form : @xmath166 in this expression @xmath167 is a constant of electron - phonon interaction , @xmath168 is a cyclic frequency of phonon , @xmath169 is a surface density of material , coefficient @xmath170 corresponds to a potential of deformation , it depends on type of phonons ( we taking in account only nonpolar acoustical and optical phonons ) , @xmath171 is a phonon distribution function , @xmath172 is an energy of electron ( in our case it described by ( [ eq : eq2 ] ) ) . because of we are interesting in motion of electrons under the influence of comparatively weak electric fields and temperatures , we shall assume phonon distribution to be close to equilibrium , so @xmath171 can be described by plank function : @xmath173 .
note , that similar to ( [ eq : eq64 ] ) expression of charge carriers probability on nonpolar phonons in two - dimensional material with linear on quasi - momentum term in energy spectrum was derived in @xcite ( for the case of graphene ) . at temperatures
less or equal to @xmath174 an acoustical phonon energy is much less than representative energies of electron in initial and finishing states and energy of thermal motion : @xmath175 .
so in arguments of dirac delta - function in ( [ eq : eq64 ] ) we can neglect a phonon energy as well as we can reduce phonon distribution function to a form @xmath176 . in a case of charge carriers scattering on long - wavelength nonpolar acoustical phonons a coefficient @xmath177 , where @xmath178 is a constant of deformation potential which typical value is about several ev .
since in long - wavelength approximation we deal with longitudinal lattice oscillations , we can represent cyclic frequency of acoustical phonons in a form @xmath179 , where @xmath137 is a speed of sound . in a such a way constant of electron - phonon interaction
is @xmath180 . consider a unit as negligible compared to @xmath181 in the second term of ( [ eq : eq64 ] ) , we obtain an expression for the probability of electron scattering by acoustic phonons in the following form : @xmath182 to find a total probability @xmath76 of electron scattering by phonons with the momentum of @xmath77 on one of the acoustic branches one should integrate ( [ eq : eq65 ] ) over the final momentum @xmath105 , as well as the configuration space : @xmath183 going over in a polar coordinate system in momentum space and substituting the explicit expression of the energy spectrum ( [ eq : eq2 ] ) , we obtain the following expression for the total probability : @xmath184 using the rule ( see , for example , @xcite ) @xmath185=\dfrac{1}{|\partial f/\partial x|}\delta(x - x_0),\ ] ] where @xmath186 is defined by condition @xmath187 , one obtains @xmath188 the final expression for the total probability of electron scattering by acoustic phonons becomes a form : @xmath189 the scattering of electrons by acoustic phonons is assumed to be elastic , i. e. the absolute value of the momentum is unchanged . during the modelling the recalculation of projections of momentum is performed , for which purpose the values of an angle @xmath163 are raffled off .
it is convenient to introduce dimensionless variables : @xmath190 , which we use already in section [ sec:2 ] .
the probability of charge carrier scattering by acoustic phonons can be presented as @xmath191 let us to write down the expression for an electron transition probability @xmath192 from the state with momentum @xmath77 to the state with momentum @xmath105 under the influence of the scattering on optical lattice oscillations . for simplicity
we shall assume phonons to be dispersionless with energy @xmath193 .
consider the case when the condition @xmath194 is satisfied .
for many materials this condition holds well at temperature about 100 k @xcite . in this case
one obtains @xmath195 , and only the second term in ( [ eq : eq64 ] ) , which describes the scattering of electron with the phonon emission , is substantial . in approximation of a deformation potential for optical lattice vibrations the coefficient @xmath196 , @xmath197 ,
so a constant of electron - phonon interaction is @xmath198 .
thus , the expression for the probability @xmath199 takes a form : @xmath200 as in the case of carrier scattering by acoustic phonons , in order to calculate the total scattering probability @xmath201 of electron with the momentum @xmath202 on the optical lattice vibrations we integrate an expression ( [ eq : eq72 ] ) on all final momentum values @xmath161 and on configuration space . @xmath203
using the rule ( [ eq : eq68 ] ) , one transforms ( [ eq : eq73 ] ) : @xmath204 where @xmath205 is a solution of equation @xmath206 @xmath207 the value @xmath205 must be non - negative real number , so one should choose @xmath208 .
the condition @xmath209 can be converted to a form @xmath210 the expression ( [ eq : eq77 ] ) has a simple physical meaning : the scattering of an electron with emission of phonon , which is taken into account in ( [ eq : eq72 ] ) , is possible if the energy of the electron is not less than the phonon energy .
passing to the non - dimensional quasi - momentum in the same way as it was done in the case of scattering by acoustic vibrations @xmath211 , we obtain an expression for the scattering probability of electron with momentum @xmath77 on optical phonons in the following form : @xmath212 here @xmath213 is an absolute value of quasi - momentum of electron after scattering by optical phonon , @xmath214 , @xmath215 , @xmath216
for a correct use of statistical methods one needs to specify the initial distribution of quasi - momenta . as in a sect .
[ sec:2 ] , we shall assume electron gas is non - degenerated so its distribution function has a boltzmann form ( [ eq : eq9 ] ) .
a probability that electron is localized in an element of quasi - momentum space @xmath217 can be presented in a form @xmath218 .
passing into a polar coordinate system one obtains @xmath219 where @xmath27 is described by ( [ eq : eq10 ] ) . going to dimensionless quantities @xmath220 and having regard to a factor @xmath221 in summation on quasi - momentum space , after integration on polar angle
one obtains @xmath222 so a probability density of distribution on quasi - momentum absolute value @xmath223 to obtain a random variable having a distribution function ( [ eq : eq82 ] ) , we use the von neumann s method ( see , for example , @xcite ) . the expression ( [ eq : eq82 ] ) possesses a maximum value at @xmath224 and this value is @xmath225 to get a needed distribution of @xmath50 one should carry out the next procedure .
two random numbers are raffled out : a first number @xmath50 is a quasi - momentum absolute value , it should be a uniformly distributed random number from an interval @xmath226 $ ] , and a second number @xmath227 is a uniformly distributed random value from an interval @xmath228 $ ] . if @xmath229 , we have found needed value @xmath50 , else we should raffle out quantities @xmath50 and @xmath227 once again .
values @xmath50 derived by this means should fill an area under the graphic of @xmath230 and depict an initial distribution of a quasi - momentum absolute value . during the modelling
we should each time after collision define a quasi - momentum absolute value and angle between quasi - momentum vector before and after collision .
expression ( [ eq : eq65 ] ) , ( [ eq : eq72 ] ) describe probability of scattering of electron on acoustical and optical phonons , respectively .
last terms in these expressions are the laws of conservation of energy which are used for determination of an absolute value of the electron quasi - momentum after scattering ( in a case of acoustical phonons this value is assumed to be equal to initial quasi - momentum of electron , in a case of optical phonons this value is defined by ( [ eq : eq79 ] ) ) .
expressions ( [ eq : eq65 ] ) , ( [ eq : eq72 ] ) are derived in approximation of long - wavelength phonons so phonon wave vector @xmath147 should satisfy the condition @xmath231 where @xmath232 is a lattice spacing .
absolute values of momentum of electron in initial and final states and momentum of phonon are the quantities of the same order .
the condition ( [ eq : eq85 ] ) for dimensionless quasi - momentum takes a form @xmath233 at temperatures @xmath234 , electron mass in order of free electron mass this inequality is well satisfied throughout the region of interest of momenta . a polar angle , which corresponds to the electron after collision , can be found based upon the fact that first terms in ( [ eq : eq65 ] ) , ( [ eq : eq72 ] ) can be treated as a density of probability of electron scattering by polar angle @xmath163 :
@xmath235 so to define an angle between the quasi - momentum vectors we use von neumann s method : we raffle out two quantities : @xmath163 which is uniformly distributed in the interval @xmath236 , and @xmath237 , uniformly distributed in the interval @xmath238 . if @xmath239 , we have found desired value @xmath163 , else we should raffle out these values again .
first of all , it is interesting to calculate the mean free time of charge carriers and find its dependence on the electric field , type and energy of phonons .
a presence of only acoustical phonons in system ca nt lead to an appearance of transverse direct current , which confirms the results of @xcite , obtained for the same problem in graphene . about the absence of the effect we conclude by the violent increasing of a direct current density variations at small values of optical phonon energy .
the effect do nt occur at @xmath240 , which is physically reasonable , because for a transverse direct current appearance an energy spectrum non - additivity is needed . on figure [ fig:5 ] there
is presented a dependence of a mean relaxation time @xmath114 on an energy of optical phonons ( we use non - dimensional quantity @xmath241 ) .
one can see that mean relaxation time is about @xmath242 and weakly increase with phonon energy rise .
the expression of probability of charge carriers scattering ( [ eq : eq78 ] ) is derived in condition of @xmath243 , so we should use as much as possible values of the ratio of the phonon energy to temperature .
the value of mean relaxation time @xmath244 , needed to effect appearance , is stringent condition for two - dimensional materials , so we shall use in further investigations @xmath245 , which at @xmath246 is equivalent to the optical phonon energy @xmath247 . ) ] on figure [ fig:6 ] there are the dependencies of the mean relaxation time @xmath114 on the constant component of electric field @xmath74 .
one can see that @xmath114 very weakly depends on the components of electric field strength of elliptically polarized wave @xmath53 and @xmath57 . with a rise of a constant component of electric field @xmath74 the mean relaxation time slowly decrease .
graphics on fig .
[ fig:5 ] and [ fig:6 ] show that approach on constant collision frequency @xmath248 , which is used in section [ sec:2 ] , is a good approximation for investigated problem .
, blue curve correspond to @xmath249 , black curve correspond to @xmath250 on a figure [ fig:7 ] the dependence of the transverse direct current density @xmath65 on @xmath74 at different values of @xmath53 and @xmath57 is presented .
solid lines on the figure correspond to the calculations on the base of the constant collision frequency approximation , points correspond to results of monte carlo simulations .
all graphics , respective to the constant collision frequency approximation , are plotted at @xmath251 , so the best coincidence of results , obtained by different methods , one can see at @xmath252 . at these values of @xmath74 , according to figure [ fig:5 ] , mean collision frequency nears to @xmath253 . on the strength of constant component of electric field @xmath74 .
red curve and points correspond to @xmath254 , blue curve and points correspond to @xmath249 , black curve and points correspond to @xmath250 the dependence of a transverse direct current density @xmath65 on phase shift @xmath61 between components of the elliptically polarized wave is presented on a figure [ fig:8 ] .
solid lines correspond to the approach of a constant collision frequency @xmath255 , points to the monte carlo simulations .
one can see , that in approach of a constant collision frequency with a high degree of accuracy @xmath256 .
same phase shift dependence is a feature of the effect of the direct current generation in transverse direction to a vector of a constant component of electric field strength under the influence of elliptically polarized electromagnetic wave which is incident normally to the surface of the sample of material with non - additive spectrum .
this result coincide with them obtained in @xcite , @xcite for the case of graphene and graphene superlattice , respectively , using the same approximation .
monte carlo simulations gives to another type of dependence : @xmath257 .
analogous result was obtain for the case of graphene in @xcite where there is a macroscopic explanation of this phase shift dependence . on the phase shift @xmath61 between the components of elliptically polarized wave .
graphics correspond to @xmath258 the modelling was carried out at next values of parameters : a constant of deformation potential of acoustic phonons is @xmath259 , a constant of deformation potential of optical phonons is @xmath260 , a mass of charge carriers is equal to the mass of free electron , a surface density of the sample is @xmath261 , a speed of sound is @xmath262 , a surface concentration of charge carriers is @xmath263 , a temperature is @xmath246 .
an energy of an optical phonon in simulations , results of which is presented on figures [ fig:6 ] , [ fig:7 ] and [ fig:7 ] , is @xmath264 .
a cyclic frequency of elliptically polarized wave is @xmath265 . at these values coefficients
@xmath266 , so differences in frequency of a charge carriers scattering on acoustical and optical phonons are defined only owing to the presence of minimal energy , required an electron to emit the optical phonons . as shown by simulation , scattering of the electron on an acoustical phonon
takes place five times more often then scattering on optical phonon .
modelling was carried out on time interval with length of 2000 in units @xmath267 , time step was @xmath268 , minimal number of particles 1000 ( in cases when we get great dispersion , we repeat our calculations several times and average out values of direct current density on these iterations ) . a typical number of collisions on acoustical phonons
was about 14000 .
motion equations [ eq : eq6 ] are solved by the runge - kutta method of the fourth order .
initial quasi - momentum distribution @xmath269 is chosen by mean , presented in section [ sec:5 ] .
our calculations show that initial distribution has a significant impact only for collisionless kinetic processes and in a case of investigations of different types of scattering it exerts a weak influence on the modelling result . in our modelling
we use parallel calculation on gpu ( graphics processing unit ) on the base of nvidia cuda technology .
we have realized a xorshift random generator @xcite in kernel function .
each generator is initialized accordingly of thread and block indexes of kernel . using own random generator in each kernel allows us to perform all calculations for each particle ( which is specified by initial conditions ) on gpu .
we find an average electron velocity on ensemble as an arithmetic mean of the velocity values in each time moment , calculated by different kernels
. then we average electron velocity on time .
for work with the gpu we used python and its package pycuda @xcite .
using the monte carlo modelling allows us to calculate a mean relaxation time , to find its dependencies on electric field components and on an energy of optical phonon . using monte carlo simulations we could directly verify the results of the approach of a constant collision frequency .
these calculations allow us to define more exactly the dependence of direct current on phase shift between the components of elliptically polarized wave . in our investigations
we examine values of electric field strengths order of tens of sgse units , which is , strictly speaking , great values .
the reason of this choice is large value of dispersion of direct current in weak electric fields , if we use monte carlo method for these calculations .
method of a constant collision frequency , used in section [ sec:2 ] , gives good results in a case of weak fields .
furthermore , this method can be used in case when @xmath270 .
on the base of approach of a constant collision frequency and using semiclassical monte carlo simulations the effect of the appearance of transverse direct current in material with a rashba hamiltonian under the influence of an elliptically polarized wave and a constant electric field is studied .
modelling on the base of monte carlo method allows us to define a mean relaxation time and its dependence on components of the electric field and on the energy of optical phonons .
this modelling shows that the conditions of application of a constant relaxation time approximation are satisfactorily hold true in the considered problem .
a comparison of direct current component calculations on the base of these two approach is carried out .
the work is supported by the grant of rfbr no .
13 - 02 - 97033 r_povolzhe_a , and it is carried out under the financial support of the ministry of education and science of the russian federation for perform public works in the field of scientific activities of the base portion of the state task number 2014/411 ( project code : 522 ) .
the work is performed using the nvidia kepler coprocessor , acquired within the framework of the program of a strategical development of the volgograd state technical university .
shorokhov , a.v . ,
khvastunov , n.n . ,
hyart , t. , alekseev , k.n . :
generation of direct current in a semiconductor superlattice under the action of a bichromatic field as a parametric effect , j. experimental and theoretical phys . ,
111(5 ) , 822829 ( 2010 ) karch , j. , olbrich , p. , schmalzbauer , m. , brinsteiner , c. , wurstbauer , u. , glazov , m.m . ,
tarasenko , s.a . ,
ivchenko , e.l . ,
weiss , d. , eroms , j. , ganichev , s.d .
: photon helicity driven electric currents in graphene , arxiv:1002.1047v1 ( 2010 ) kryuchkov , s.v . ,
kukhar , e.i . : mutual rectification of cnoidal and sinusoidal electromagnetic waves with orthogonal polarization planes in a graphene - based superlattice , optics and spectroscopy , 112(6 ) , 914919 ( 2012 ) zavyalov , d.v . , kryuchkov , s.v .
, kukhar , e.i . : effect of transverse entrainment of charge carriers by the field of two electromagnetic waves in a semiconductor , phys . of the solid state , 54(9 ) ,
1853-1856 ( 2012 ) gusev , g.m . ,
kvon , z.d . ,
magarill , l.i . ,
palkin , a.m. , sozinov , v.i . ,
shegai , o.a . ,
entin , m.v . : resonant photovoltaic effect in an inversion layer at the surface of a semiconductor , jetp lett . , 46(1 ) , 3336 ( 1987 ) zhang , q. , wang , x.q .
, yin , c.m . , xu , f.j . , tang , n. , shen , b. , chen , y.h . ,
chang , k. , ge , w.k . , ishitani , y. , yoshikawa , a. : strong circular photogalvanic effect in zno epitaxial films , appl .
97 , 041907 ( 2010 ) yu , j.l . ,
chen , y.h .
, liu , y. , jiang , c.y . , ma , h. , zhu , l.p .
: spectra of rashba- and dresselhaus - type circular photogalvanic effect at interband excitation in gaas / algaas quantum wells and their behaviors under external strain , appl .
, 100 , 152110 ( 2012 ) ganichev , s.d . ,
belkov , v.v . ,
schneider , p. , ivchenko , e.l . ,
tarasenko , s.a . ,
wegscheider , w. , weiss , d. , schuh , d. , beregulin , e.v . ,
prettl , w. : resonant inversion of the circular photogalvanic effect in n - doped quantum wells , phys .
b , 68 , 035319 ( 2003 ) giglberger , s. , golub , l.e . , belkov , v.v . ,
danilov , s.n .
, schuh , d. , gerl , c. , rohlng , f. , stahl , j. , wegscheider , w. , weiss , d. , prettl , w. , ganichev , s.d . :
rashba and dresselhaus spin splittings in semiconductor quantum wells measured by spin photocurrents , phys . rev .
b , 75 , 035327 ( 2007 ) rashba , e.i . : properties of semiconductors with an extremum loop .
1 . cyclotron and combinational resonance in a magnetic field perpendicular to the plane of the loop , sov .
solid state , 2 , 1109 ( 1960 ) tarasenko , s.a . ,
ivchenko , e.l .
, belkov , v.v . , ganichev , s.d . ,
schowalter , d. , schneider , p. , sollinger , m. , prettl , w. , ustinov , v.m . ,
zhukov , a.e . ,
vorobjev , l.e . : monopolar optical orientation of electronic spins in semiconductors , arxiv : cond - mat/0301393 ( 2003 ) cartoixa , x. , ting , d.z .- y . ,
mcgill , t.c . :
theoretical investigations of spin splittings and optimization of the rashba coefficient in asymmetric alsb / inas / gasb heterostructures , j. comput . electron .
, 1 , 141146 ( 2002 ) | an appearance of a direct current perpendicularly to a constant component of an electric field in material with rashba hamiltonian under the influence of an elliptically polarized wave is investigated in one - subband approximation . on its physical nature
this effect is close to a circular photogalvanic effect ( cpge ) on intraband transitions .
the effect is studied on the base of two approaches : investigations of boltzmann kinetic equation in a constant collision frequency approximation and semiclassical monte carlo simulations , which immediately takes into account microscopic processes of charge carriers scattering on optical and acoustical phonons .
monte carlo modelling allows us to determine a mean relaxation time and its dependencies on electric field strengths and on the energy of optical phonons . on the base of these estimations a possibility of using a constant relaxation time approximation
is justified .
it is confirmed that the main contribution to the effects of transverse rectification is made by inelastic scattering of electrons on optical phonons .
a comparison of results of monte carlo simulations and calculations on the base of a constant collision frequency approximation is presented .
example.eps !
ps - adobe-3.0 epsf-3.0 gsave newpath 20 20 moveto 20 220 lineto 220 220 lineto 220 20 lineto closepath 2 setlinewidth gsave .4 setgray fill grestore stroke grestore |
in recent years , a great deal of attention has been devoted to magnetic organometallic compounds due to their prospective use as nanoscale magnetic devices . for example , single - molecule magnets ( smms)@xcite are currently considered as potential building blocks for spintronics,@xcite molecular electronics,@xcite and quantum computing.@xcite magnetic organometallic molecules can be formed by bringing together small clusters of transition metal atoms and benzene molecules , such as in the case of cobaltocene and ferrocene .
magnetic molecular clusters have been synthesized recently by molecular beams.@xcite their magnetic moments have been measured using stern - gerlach molecular - beam deflections showing that part of the original magnetization may survive when the number of transition metal centers and ligands are comparable .
@xcite contrary to the smms , where the transition metal centers are surrounded by organic compounds and thus , protecting their magnetic moment,@xcite the transition metal atoms in such organometallic clusters are in close proximity , leading to strong exchange interactions .
although these organometallic clusters made up of transition metal clusters and benzene rings share same magnetic properties with smms , their magnetism is much less studied .
bz@xmath0 cluster deposited on a substrate .
lower panel , most stable configuration of the cluster .
the cobalt atoms ( light blue ) form an equilateral triangle , and the benzene rings ( carbon shown in yellow and hydrogen in cyan ) are perpendicular to the plane formed by the cobalt atoms .
, scaledwidth=45.0% ] the prototype organometallic clusters were based on cobalt because such clusters have predominantly ferromagnetic orders and typically have high magnetic anisotropies .
experimental results suggest that clusters of late transition metals fully saturated with benzene molecules are stable @xmath1-complexes that are not very reactive.@xcite rather than having the metal center(s ) sandwiched between benzene rings , as was typical in early transition metal complexes,@xcite the cobalt centers in cobalt - benzene ( co - bz ) clusters are close together , forming an inner metallic cluster bound through @xmath1-bonds to benzene molecules around the cluster , commonly called rice - ball configuration .
it appears that magnetic order is retained in these small co - bz clusters .
co - bz complexes have been subjected to molecular beam deflection magnetic studies between @xmath2 k and @xmath3 k , with their magnetic moments extracted by fitting the experimental data to the curie law , displaying a superparamagnetic behavior.@xcite the magnetic moments per cobalt atom of co@xmath4bz@xmath5 clusters with @xmath6 , have been found to range from @xmath7 to @xmath8 and to decrease as the cluster size increases .
@xcite these magnetic moments were all higher than the bulk cobalt magnetic moment of @xmath9 . however , marked discrepancies between magnetic moments determined in different experiments have been reported.@xcite henceforth , the magnetic properties of co - bz clusters and their response to magnetic fields require further research . thus , the aim of this work is to theoretically examine these properties taking the meaningful example of co@xmath0bz@xmath0 cluster .
we consider co@xmath0bz@xmath0 due to its small size and its closed structure .
it should be noted that co@xmath0bz@xmath0 clusters have been found to be more stable and abundant than other sizes , a fact that can be analyzed by mass spectrometry.@xcite in our study we explore both the collinear and non - collinear magnetic orders .
we used several density functional theory ( dft ) approaches . for our most accurate and detailed description
, we used a relativistic and full - potential linearised augmented - plane wave ( fp - lapw ) @xcite including the effects of spin - orbit interactions .
our results show that magnetic coupling between the cobalt atoms is strongly modified by the benzene molecules . and associated to its closed geometry , which lead to the existence of non - trivial magnetic orders and geometrical spin frustration .
crucially , we found that the minimum energy configuration corresponds to a non - collinear order in which the magnetic moments lie in the plane defined by the three cobalt atoms , which is almost @xmath10 mev below the first excited ferromagnetic configuration .
finally , based on the results of the dft calculations , we introduced a minimum quantum - spin model that accounts for both , magnetic anisotropies and inter - spin couplings .
the results of the model suggest that the most stable configurations corresponds to a non - symmetric exchange interaction between the three cobalt atoms with a significant local anisotropy due to the benzene rings .
we investigated the relaxed structures using electronic structure calculations resolved for spin within the dft framework .
the kohn - sham ansatz uses a linear combination of atomic orbitals , as used in the siesta code.@xcite the relaxed structures produced using siesta were successfully tested and refined using two other plane - wave codes , vasp@xcite and quantum espresso.@xcite we started with different input geometries and determined how the magnetic properties changed . in general , describing the magnetic properties of systems including transition metals in detail requires a proper relativistic treatment of electron - nuclear and electron - electron interactions.@xcite we deal with spin - orbit interactions using an accurate relativistic version of the full - potential linearized augmented plane - wave method@xcite as it is implemented in the elk - code.@xcite a more detailed information on the dft calculations is provided in the supplemental material .
we first consider the co@xmath0bz@xmath0 cluster with collinear magnetic moments .
we optimized the structures using siesta , then performed an exhaustive search to determine the ground state by exploring the different relative orientations of the benzene rings .
the geometric relaxations lead to the structure shown in fig .
[ fig1:esq ] .
the ground state was found to be antiferromagnetic ( afm ) with a total magnetic moment of @xmath11 , each cobalt atom contributing @xmath12 .
the ground state corresponds to a structure with the cobalt atoms arranged in a slightly distorted equilateral triangle , with two bonds lengths of d@xmath13 = d@xmath14 = @xmath15 and one of d@xmath16 = @xmath17 .
the cobalt atoms were situated in hollow positions with respect to the closest benzene ring , @xmath18 above the hexagon barycentre
. there were two higher energy ferromagnetic ( fm ) states , with total moments of 5 @xmath19 and 3 @xmath19 , at @xmath20 and @xmath21 ev , respectively .
these excited ferromagnetic configurations have been identified as the as the ground state in previous calculations.@xcite note that when repeating the collinear calculations with quantum espresso and vasp , we found similar afm configurations as ground state .
the calculated co@xmath0bz@xmath0 ground state band gap was @xmath22 ev , and probably even larger because it is well known that band gaps calculated using the generalized gradient approximation are greatly underestimated.@xcite the large band gap indicates that co@xmath0bz@xmath0 is very stable , in good agreement with the experimental results.@xcite we compared the results described above with the results of bare co@xmath0 cluster .
we found that the bare co@xmath0 cluster formed a ferromagnetic isosceles triangle with two bond lengths of @xmath23 and one of @xmath24 .
this can be ascribed to the jahn - teller effect distorting the bare co@xmath0 cluster.@xcite the asymmetry in this structure was found to be reflected in the local magnetic moments , @xmath25 , @xmath26 , and @xmath26 , giving a total magnetic moment of @xmath27 @xmath19 .
this fm configuration was clearly preferred over the afm configuration with an energy difference of @xmath28 ev.@xcite these results were in clear contrast with the results for co@xmath0bz@xmath0 .
adding benzene molecules appears to almost remove the distortion of the co@xmath0 cluster .
this could be partly due to charge transfer , because cobalt atoms gain about half an electron from the benzene molecules . to verify this hypothesis , we performed calculations for the charged co@xmath29 cluster ground state and found that structural relaxation causes the cluster to tends toward an equilateral geometry .
the red arrows point the directions of the local magnetic moments of the cobalt atoms .
each label shows the energy difference ( in mev ) between the configuration shown and the ground state . in parentheses ,
the magnitude of the local magnetic moment at each cobalt atom ( in units of @xmath19 ) is given in the atom order shown in fig .
[ fig1:esq ] .
note that the non - collinear configurations become the ground state .
, scaledwidth=48.0% ] the antiferromagnetic ground state identified in the collinear calculations described in sec .
[ sec : relaxation ] is certainly peculiar to co@xmath0bz@xmath0 . the co@xmath0 cluster in particular and the nanostructures based in pure cobalt in general are ferromagnetic .
antiferromagnetic interactions and non - collinear magnetic landscapes may naturally give rise to geometrical frustration as it is found in spin glasses.@xcite with these ideas in mind , we performed dft calculations using a non - collinear full - potential linearized augmented plane - wave method for the previously described fully relaxed co@xmath0bz@xmath0 cluster .
we found an extra set of non - collinear states with in - plane magnetization , with the new ground state @xmath30 mev lower in energy that the collinear afm configuration .
the magnetic configurations could be divided into three groups according to the total energies and the orientation of the magnetizations of the cobalt atoms , as shown in fig .
[ fig2:scale ] .
the first group contains low energy states with in - plane magnetic moments , aligned perpendicular or parallel to the benzene rings .
the second group , with energies approximately @xmath31 mev higher than the ground state energy , has radial magnetization pointing either towards or away from the centroid of the cobalt atoms triangle . in both groups
the magnetic moment per co atom is between @xmath32 and @xmath33 .
a third group is found with collinear states with magnetizations perpendicular to the cobalt atoms plane . the afm configuration described in sec .
[ sec : relaxation ] lies @xmath34 mev above the ground state , and the first fm state with a magnetic moment of @xmath35 was found at @xmath36 mev above the ground state .
it is noteworthy that in these cases , it was necessary to fix the direction of magnetization .
starting with a collinear configuration without fixing the directions of the magnetic moments led to the electronic structure calculations converging on non - collinear arrangements because of strong spin - orbit coupling of cobalt atoms.@xcite the ground state magnetization density in the cobalt plane is shown in fig .
[ fig3:map ] .
the magnetic moments are highly localized around the cobalt atoms .
this strongly supports the idea of local moments used in the heisenberg - like model described in sec .
[ sec : modelh ] .
in addition , the direction of the magnetization density rotates in the plane of the cobalt atoms generating a vortex - like pattern .
similar observations have been made for triangular and kagome transition - metal networks.@xcite we tested the effects of stronger interactions between the 3d electrons in the cobalt atoms by adding the hubbard @xmath37 and @xmath38 terms .
we used typical values of @xmath39 ev for on - site coulomb repulsion and @xmath40 ev for the exchange.@xcite the magnetic orders are qualitatively the same as the results shown in fig .
[ fig2:scale ] and they still are separable into the same three groups . in particular , the ferromagnetic configuration is now at @xmath41 mev , while the antiferromagnetic is at @xmath42 mev above the ground state . as a result of the hubbard interaction ,
the localization of the d - states increased with local magnetic moments slightly enhanced .
bz@xmath0 ground state .
the small arrows point in the direction of the local magnetic moment , and the colors indicate the magnitudes of the vectors . each green circle indicates the muffin - tin radius of a cobalt atom , and the large arrow in each green circle indicates the total magnetic moment assigned to the cobalt atom .
, scaledwidth=50.0% ]
a detailed description of the magnetization dynamics and the temperature dependence requires to take into account excited states .
therefore , in this section we introduce a minimum quantum spin model capable of describing such magnetic properties .
the proposed hamiltonian has the form @xmath43 where @xmath38 and @xmath44 are exchange constants , @xmath45 is the spin operator corresponding to the @xmath46-cobalt atom ( fig .
[ fig1:esq ] ) , and @xmath47 is the local magnetic anisotropy of this atom induced by the benzene rings . the hamiltonian shown in eq .
( [ hheis ] ) implies two basic assumptions .
the first is that the spin of each cobalt atom is quantized and the magnitude of @xmath48 is conserved .
this condition was fully satisfied for the low energy configurations shown in fig .
[ fig2:scale ] .
the second assumption is that the magnetic moments are localized on the cobalt atoms .
this was supported by the magnetic solutions provided by the dft calculations , as shown in fig .
[ fig3:map ] .
moreover , the presence of two different exchange couplings @xmath38 and @xmath44 is compatible with the _ isosceles_-type interactions indicated by the electronic structure calculations . indicates the limit of _ equilateral _ interactions . ] the low energy configurations with non - radial symmetry ( fig .
[ fig2:scale ] ) indicate @xmath49.@xcite here we take a spin @xmath50 for each cobalt atom .. for comparison , we also used a spin @xmath51 , without finding significant qualitative differences .
a more detailed discussion of the spin value is given in the supplemental material . ] for the local magnetic anisotropy we use the lowest order terms compatible with the @xmath52 symmetry , @xmath53,@xcite where @xmath54 is the uniaxial anisotropy parameter along the ( local ) high symmetry axis @xmath55 .
these symmetry arguments are in good agreement with the dft results for the cobz cluster .
the total hamiltonian , which needs all the spin operators in eq .
( [ hheis ] ) to be written in a common reference frame , was solved through the numerical diagonalization of the @xmath56 hamiltonian matrix .
we shall denote the eigenvalues and eigenvectors of @xmath57 by @xmath58 and @xmath59 respectively .
the phase space of the model given in eq .
( [ hheis ] ) was quite rich , as described in more detail in the supplemental material .
we therefore searched for the parameters @xmath38 , @xmath44 and @xmath54 that better reproduced the main features of the dft total energies .
these features were in order of increasing energy , as shown in fig [ fig2:scale ] : ( i ) a ground state with magnetization in the cobalt plane , ( ii ) a first excitation at very low energies , @xmath60 mev , with in - plane magnetization , ( iii ) a second excitation at @xmath61 mev , ( iv ) and a first out - of - plane magnetization excitation @xmath62 mev with antiferromagnetic order .
we found two possible scenarios with asymmetric interactions : @xmath63 ( weak / strong / strong co - co interaction ) , and with @xmath64 ( weak / weak / strong co - co interactions ) .
the best overall agreement according to the points mentioned above is found for the asymmetric configuration with @xmath65 and @xmath66 .
nevertheless , a configuration with @xmath67 and @xmath68 can not be fully discarded , so both are expected to be experimentally relevant under certain conditions .
this kind of duality has also been found experimentally in magnetic clusters containing triangular mn@xmath0 clusters.@xcite spin hamiltonians are frequently used to connect experimentally observable quantities with the microscopic variables.@xcite for instance , a spin hamiltonian similar to eq .
( [ hheis ] ) was fitted to experimental data for a supramolecular aggregate containing smms , finding that it successfully explained the observed complex magnetic susceptibility and relaxation times behaviors.@xcite accurate spectral information on the magnetic excitation of the clusters can be extracted using various techniques , including inelastic neutron scattering , optical spectroscopic methods , magnetic susceptibility measurements and electron paramagnetic resonance ( epr).@xcite due to the energy scales involved in the problem , few of these techniques could provide meaningful magnetic information .
the analysis of these characteristic scales lead us to consider epr ( see details in the supplemental material ) . by using perturbation theory in the _ ac _ field
, one can arrive to the following expression for the amplitude of the epr signal @xmath69 where @xmath70 denotes the thermal equilibrium occupation of the @xmath71-energy level of eq .
( [ hheis ] ) , and @xmath72 with @xmath73 the total spin .
all the frequency dependence was included in the factor @xmath74 .
a proper theoretical description of the epr line - shape requires to account for the relaxation and decoherence mechanisms , which is beyond the scope of the present work .
instead , we used the frequently found experimental line - shape of @xmath75 , which can be approximated by the sum of two lorentzians centred at @xmath76 and with half - width at half - maximum of @xmath77 , the transversal relaxation rate.@xcite for a small frequency window the resonant states have a non - negligible contribution .
these states have bohr transition frequencies satisfying @xmath78 . for low enough temperature only the ground state @xmath79 is occupied .
the epr signal is thus significant for the states with @xmath71 such that @xmath80 .
k , the first excited state can have a significant population at @xmath81 t. thus , there may be a transition between the first and second excited states with a frequency around @xmath82 ghz .
in the case of diluted clusters , one might expect @xmath83 ms , so resonant peaks will not overlap . ] the angular dependence of epr signal should reflect the magnetic anisotropy of the cluster .
here we focused on the transition between the ground state and the first excited state , although the excitation to the second excited state may be also probed .
we assumed a static _ dc _ field applied forming a polar angle @xmath84 with the normal @xmath85 to the cobalt plane , an azimuthal angle @xmath86 with the @xmath87 vector , see inset in fig .
[ figs4 ] , and a small _ ac _ field perpendicular to @xmath88 .
figure [ figs4 ] shows the angular dependence of the _ on - resonant signal _ @xmath89 .
it is noteworthy that the angular dependence enters in the resonant signal on two counts .
first , due to the magnetic anisotropy of the cluster , the excitation energy changes with the angle as reflected in the occupation difference @xmath90 .
this variation was observed in the signal dependence with temperature ( fig .
[ fig4 m ] of supplemental material ) .
k since above the broadening @xmath77 can display a significant temperature variation .
] second , the oscillator strength of the transition changes through a variation of the @xmath91 matrix elements .
we found that the resonant signal shows maxima when the _ dc _ field is applied perpendicular to the cobalt atoms plane ( @xmath92 ) , and an _ ac _ field pointing along the easy axis of atom 1 through coupling @xmath38 in eq .
( [ hheis ] ) .
the resonant signal is more dramatically modulated in the @xmath86-angle .
these angular dependences should be observable in epr experiments .
, versus the polar @xmath84 and azimuthal @xmath86 angles ( see lower inset ) .
the ac field direction is parallel to @xmath93 .
the hamiltonian parameters were taken for the optimal agreement with the dft results and the temperature was fixed to 1 k. , scaledwidth=50.0% ]
we investigated the prototype co@xmath0bz@xmath0 molecular cluster using several dft methods ( the siesta , vasp , quantum espresso , and elk methods .
we found that the benzene groups induce non - collinear magnetic configurations in the cobalt cluster .
we demonstrated that the ground state of the cluster corresponds to configurations in which the magnetization remains in the plane defined by the three cobalt atoms .
this is especially striking when compared with the results for a bare co@xmath0 cluster , which has a ferromagnetic ground state with out - of - plane magnetization .
we have therefore shown that the benzene rings play a crucial role in the magnetic landscape of the co@xmath0bz@xmath0 cluster .
to offer further insights into the responses of the magnetic properties , we built a quantum spin model hamiltonian that accounted for the low energy magnetic solutions of the electronic structure calculations .
this hamiltonian contains local uniaxial magnetic anisotropy , associated with the presence of the benzene rings , and antiferromagnetic coupling between the spins of the co atoms .
this approach leads to two possible stable configurations corresponding to different deviations from perfectly symmetric interactions .
in contrast to the symmetric case , which displays spin frustration in the heisenberg limit,@xcite the anisotropic and asymmetric molecule leads to an effective total spin of @xmath94 .
a detailed study of the energy scales lead us to propose epr as the most suitable technique to analyze the spectral magnetic features of the co@xmath0bz@xmath0 cluster within the reach of experimentally realizable conditions .
in addition , we demonstrated that the co@xmath0bz@xmath0 cluster has a preferential in plane anisotropy , revealed through a drastic azimuthal angular dependence between the direction of the radiofrequency field and the easy axis of each co atom .
a meticulous experimental determination of the magnetic anisotropy of the molecule would be even possible , as done for instance in single molecule tunnel junctions based on organic molecules.@xcite
this work was partially supported by the spanish ministry of economy and competitiveness mineco through projects fis2013 - 48286-c2 - 1-p ( jwg , ta - l , aa ) and mat2015 - 66888-c3 - 2-r ( fd ) , the basque government through the nanomaterials project ( grant ie14 - 393 ) under the etortek program _ nanogune14 _ ( jwg , ta - l , aa ) , and the university of the basque country ( grant no . it-366 - 07 ) ( jwg , ta - l , aa ) .
we would also like to acknowledge the assistance of the dipc computer center and the useful discussions with l. chico .
51 natexlab#1#1bibnamefont # 1#1bibfnamefont # 1#1citenamefont # 1#1url # 1`#1`urlprefix[2]#2 [ 2][]#2 , * * , ( ) .
, , , , , , , , , , , * * , ( ) . , , , , , , , * * , ( ) . , , , , , , * * , ( ) . , * * , ( ) . , * * , ( ) . , , , , , , , , , , , * * , ( ) . , , , * * , ( ) . , * * , ( ) . , , , , , * * , ( ) . , , , , , * * , ( ) . , * * , ( ) . , , , , , , , , * * , ( ) . , , , , , , , * * , ( ) . , , , , * * , ( ) . , , , , , , , , * * , ( ) . , , , , * * , ( ) . , * * , ( ) . , _ _ ( , ) . , , , * * , ( ) . , , , , , , , * * , ( ) . , * * , ( ) . , * * , ( ) . , , , , , , , , , , , * * , ( ) , http://www.quantum-espresso.org . , * * , ( ) . _ _ , , . , * * , ( ) . , * * , ( ) . , * * , ( ) . , , , _ _ ( , , ) . , , , , , * * , ( ) . , pp . ( ) . , , , , , , , ,
* * , ( ) . , , , * * , ( ) . ,
* * , ( ) .
, , , * * , ( ) . , , , ,
* * , ( ) . , , , , , , ( ) . ,
_ _ ( , ) . , * * , ( ) . , , , * * , ( ) . , , , , , , , , * * , ( ) . , , , , , , , , , ,
* * , ( ) . , , , * * , ( ) . ,
* * , ( ) . ,
* * , ( ) . , , , * * , ( ) . , * * , ( ) . , , , , ( ) . , ,
, * * , ( ) . , _ _ , vol .
* supplemental materials : non - collinearity in small cobalt - benzene molecular clusters *
siesta calculations were performed using the exchange and correlation potentials with the generalized gradient approximation ( gga ) in perdew - burke - ernzenhof form.@xcite we used an electronic temperature of @xmath95 mev and a mesh cutoff of @xmath96 ry for all the calculations .
the atomic cores were described using nonlocal norm - conserving relativistic troullier - martins@xcite pseudopotentials with non - linear core corrections factorized in the kleynman - bylander form .
the pseudopotentials were tested to ensure that they accurately reproduce the eigenvalues of different excited states of a bare atom .
we tested the cobalt pseudopotential against the bulk cobalt .
we use the 4s@xmath973d@xmath98 valence configuration , and found that the ground state was the hcp structure , which had a stability of @xmath99 ev / atom greater than that of the fcc structure . for the hcp structure ,
the first neighbor distance was @xmath100 and the magnetic moment per atom was @xmath101 @xmath102 , in good agreement with experimental results.@xcite the cobalt pseudopotential was described in detail and used to study co@xmath103 clusters adsorbed onto graphene in a previous study.@xcite the basis sets for cobalt , hydrogen , and carbon were the double - polarized basis sets identified in previous studies .
geometric relaxations used conjugate gradient structure optimization .
the hydrogen atoms of the benzene rings made it necessary to impose small convergence thresholds on forces of the order of @xmath104 ry / bohr @xmath105 ev / .
we found that imposing less strict convergence conditions over both the energies and the forces may erroneously lead to deformed structures . in each case , a single @xmath106-point was chosen for the calculations and @xmath107 of empty space was added to avoid interactions between nearest - neighbors cells .
planewave methods such as quantum espresso@xcite and vasp@xcite were used .
self - consistent calculations were performed very accurately using a plane - wave kinetic energy cutoff of @xmath108 ry . as in the siesta calculations ,
the generalized gradient approximation in exchange - correlation was in perdew - burke - ernzerhof ( pbe)@xcite form .
norm - conserving troullier - martins pseudopotentials@xcite were used in quantum espresso , and a projector augmented wave potential construction was used in vasp .
all the atoms were allowed to relax by conjugate gradients until the forces converged , with a tolerance of @xmath109 ev / .
we then checked that the relaxed geometries reproduced the siesta results . an all - electron full - potential linearised augmented - plane wave ( fp - lapw )
were performed using the elk code .
the electronic exchange - correlation potential was treated within the local spin density approximation ( lsda ) @xcite to avoid gradients effects of the generalized gradient approximation in the non - collinear calculations .
wavefunctions in the interstitial density and potential were expanded into plane waves with a wavevector cutoff of @xmath110 , where @xmath111 is the muffin - tin radius of cobalt .
the dft results indicated that the energy levels corresponding to the @xmath112-orbitals of the cobalt atoms in a cobz cluster split in two , with the lowest orbital triplet almost fully occupied and a half - occupied excited doublet .
thus , the filling of the levels following the hund s rule leads to a spin @xmath94 .
although the degeneracy of the lowest orbital triplet is partially broken due to the new crystal field in the co@xmath0bz@xmath0 cluster , the qualitative filling is essentially conserved .
consequently , we assumed later that each cobalt behaves as a @xmath94 spin .
this spin is also compatible with the @xmath113f ground state observed in the gas phase.@xcite in order to include the local magnetic anisotropy , we made the following symmetry considerations .
an isolated cobz unit has c@xmath114 symmetry .
the local spin hamiltonian compatible with this symmetry can be written in terms of the ( tesseral tensor ) stevens operators @xmath115 as@xcite @xmath116 where @xmath117 are real coefficients .
@xmath115 are in turn linear combinations of spin operator components .
the lowest order non - isotropic term corresponds to the operator @xmath118 , which is the uniaxial term used in this work .
higher ( even ) powers of @xmath119 do not introduce any qualitative change to the energy spectrum and they can be thought as a renormalized @xmath54-value , hence we neglected them .
all these terms commute with @xmath119 , and thus , they do not introduce mixing between the eigenstates of @xmath120 .
importantly , the lowest transverse terms compatible with the c@xmath114 symmetry of cobz involves the sixth power of the ladder operator , @xmath121 .
they therefore do not contribute to the spectrum of a spin @xmath122 , so we discarded them .
as stated in the main text , the spin hamiltonian ( [ hheis ] ) contains three fitting parameters , @xmath38 , @xmath44 and @xmath54 , for a given local spin @xmath123 . herein , we described in detail the derivation of the optimal parameters that best match the information provided by the dft calculations . the electronic structure calculations of a single cobz molecule gave an energy difference of @xmath124 mev between configurations with the magnetization out of plane ( lowest energy ) and in - plane , suggesting that the easy axis would be out of plane with @xmath125 mev for @xmath94 .
the dft results for the cobz complex show that changing the magnetization orientation along the benzene plane did not significantly change the energy of the complex , in good agreement with our symmetry arguments .
we found that the benzene rings are not deformed appreciably when a co@xmath0bz@xmath0 cluster is formed , so it may be tempting to use the local anisotropy of each cobz unit .
however , this option was strongly discouraged by the dft results .
first , the distances between the cobalt atoms ( 2.34 ) were of the same order than the distance between a cobalt atom and the closest carbon , approximately @xmath126 .
second , the charge of the cobalt ions in a cobz and the co@xmath0bz@xmath0 cluster differ by half an electron .
third , the distances between cobalt atoms and the closest benzene plane changes by as much as 9@xmath127 . * a ground state with magnetization in the cobalt atom plane , * a first excitation at @xmath128 mev , with in - plane magnetization , * a second excitation at @xmath129 mev ( in - plane magnetization ) , and * a third excitation at @xmath130 mev with out - of - plane magnetization .
our approach to set the three parameters consisted of the following steps .
first , for three different @xmath131 scenarios , we looked at the qualitative behavior of the energy spectra with the @xmath132 ratio between the limits of the isolated cobz units ( @xmath133 ) , and the isotropic cluster ( @xmath134 ) .
crucially , the magnetic configurations resulting from the non - collinear calculations indicated the dominance of spin - exchange interactions over anisotropy , so the most natural scenario corresponded to @xmath135 .
the excitation energies @xmath136 from the ground state are plotted against @xmath132 for three different @xmath131 values in fig .
[ figs1 ] .
we associated the first excitation with the dft electronic state at @xmath137 mev above the ground state .
the second energy level was ascribed to the configurations found @xmath138 mev above the ground state .
thus , we choose the parameters such that @xmath139 while keeping the gap @xmath140 as large as possible in terms of @xmath54 to avoid unrealistically large anisotropy barriers .
the third magnetic configuration at @xmath141 mev , identified from the dft results as an afm configuration with out - of - plane magnetization , was taken as the third excited state of the model . comparing the energy spectra shown in fig .
[ fig2:scale ] and fig .
[ figs1 ] , we observe that the first two conditions are quantitatively satisfied for @xmath142 in the asymmetric case with @xmath143 or for @xmath144 $ ] with @xmath65
. the symmetric configuration @xmath145 did not provide a qualitative agreement for any @xmath132 ratio .
a more quantitative picture is obtained by inspection of the excitation spectrum versus @xmath131 , shown in fig .
[ figs2 ] .
the @xmath146 and @xmath147 ratios were plotted versus @xmath131 for three different values of @xmath132 , with the ideal ratios close to 0 and 3.3 , respectively . for reference ,
the corresponding excitation spectra are plotted in grey .
the best agreement is found for @xmath67 and , especially , for @xmath64 .
again , the symmetric case ( @xmath145 ) was far from the ideal ratios for any @xmath132 in the considered range of parameters .
summing up , optimal agreement with the dft results was found for @xmath148 mev and @xmath149 mev .
these values are higher than those typically found within smms .
however , the cobalt centers in the co@xmath0bz@xmath0 cluster are much closer , so a direct exchange mechanism is possible .
thus , it seems clear that both local magnetic anisotropy due to the benzene rings and the direct - exchange between the local moments associated to the co atoms plays a crucial role . before trying to connect to an specific measurement
, it is worth looking at the energy scales involved .
if we denoted the excitation energies from the ground state to the first and second excited states as @xmath150 and @xmath140 respectively , and introducing the temperature @xmath151 , magnetic field @xmath152 , and frequency @xmath153 , we got the following energy scales at zero external field : @xmath154 mev ( @xmath155 k , @xmath156 t , and @xmath157 ghz ) , and @xmath158 mev ( @xmath159 k , @xmath160 t , and @xmath161 thz ) .
this had the following consequences for the possible experimental observation .
static measurements of the susceptibility may provide significant information about the lowest energy excitation for @xmath162 k. furthermore , for @xmath163 close to room temperature there may be other excitations not included here , like phonons .
we had not found significant changes of the susceptibility with temperature or magnetic fields below 10 t. _ ac _ measurements of the dynamical susceptibility are also commonly used to extract additional magnetic information of smms , but the frequency range is limited to @xmath164$],@xcite clearly outside the energy range of interest .
high @xmath165-field epr measurements may provide an accurate spectral information .
in fact , for the @xmath81 t @xmath38-band , the electron spin resonance is found around @xmath166 ghz.@xcite for such an applied external field applied out of the co s plane , @xmath167 mev and @xmath168 mev ( @xmath169 ghz ) , within the range of experimental frequencies .
epr experiments apply a fixed _ dc _ field along a given direction , which we defined as the quantization axis @xmath170 , and a small perpendicular _ ac _ field @xmath171 , along the @xmath172-axis . for the co - bz molecule ,
the most convenient set - up corresponds to the @xmath38-band ( @xmath81 t ) with frequencies @xmath173 ghz .
we used time dependent perturbation theory because the perpendicular _ ac _ field is much smaller , with typical intensities of the order of 1 g. the absorbed power is given as the variation of the average energy @xmath174 $ ] with time , where @xmath175 is the density matrix .
we here assumed a coherent dynamics .
we used an interaction picture with respect to the term @xmath176 and a first order perturbative expansion for the density matrix operator .
thus , the instantaneous power was given by @xmath177 , \crcr & & \label{ipower}\end{aligned}\ ] ] where @xmath178 , @xmath179 is the gyromagnetic ratio , and @xmath70 denotes the thermal equilibrium occupation of the energy level @xmath71 .
there are however two considerations to be made in order to connect the absorbed power @xmath180 with the epr signal .
first , the absorption signal of the epr measurement corresponds to the average of instantaneous power over a measurement time @xmath181 .
second , the coupling of the spin system with the environment induces a dissipative dynamics .
when two energy levels are close to resonance for a given frequency @xmath182 , i.e. , @xmath183 , the incoherent dynamics introduces two new time scales , the _ longitudinal or relaxation time _ @xmath184 and the _ transversal or decoherence time _ @xmath185.@xcite these two time scales determine the line - shape of the epr absorption spectra and , in particular , the width of the resonances , given essentially by @xmath186 . when @xmath180 is averaged over a long period of time @xmath187 , the first term cancels . for a frequency window @xmath188 , the _ steady - state _
absorbed power is then proportional to @xmath189 .
the incoherent dynamics modifies the time dependence in the last part of eq .
( [ ipower ] ) which , in the frequency domain , leads to a finite resonant amplitude and a finite frequency width .
we approximated this frequency dependence using the collision broadened profile of van vleck and weisskopf ( 1945),@xcite @xmath190 where @xmath75 takes the form @xmath191 . \nonumber\end{aligned}\ ] ] however , we should keep in mind that this epr line shape does not account for saturation effects and the associated temperature increment , which for a typical _ ac _ field of @xmath192 g occurs for longitudinal relaxation time @xmath193s .
, versus the angle @xmath84 formed by the normal to the co s plane and the _ dc _ magnetic field ( @xmath194 ) .
field is parallel to one of the easy axis of atom 1 in eq .
( [ hheis ] ) ] .
, scaledwidth=50.0% ] | cobalt clusters covered with benzene in the form of rice - ball structures have recently been synthesized using laser ablation . here , we investigate the types of magnetic order such clusters have , and whether they retain any magnetic order at all .
we use different density functional theory ( dft ) methods to study the experimentally relevant three cobalt atoms surrounded by benzene rings .
we found that the benzene rings induce a ground state with non - collinear magnetization , with the magnetic moments localized on the cobalt centers and lying on the plane formed by the three cobalt atoms .
this is surprising because nanostructures and small clusters based on pure cobalt typically have a predominantly ferromagnetic order , and additional organic ligands such as benzene tend to remove the magnetization .
we analyse the magnetism of such a cluster using an anisotropic heisenberg model where the involved parameters are obtained by a comparison with the dft results .
moreover , we propose electron paramagnetic resonance as an experimental tool to study the anisotropic response . |
a considerable fraction of chemically peculiar a and b stars ( cp stars ) show strong surface magnetic fields of about 500 g up to over 10 kg .
the evolution of these fields from star formation to their current presence is unknown .
this paper compiles a number of physical processes which may or may not play a role in the whole scenario of cp star magnetism .
the fraction of main - sequence stars with radiative envelopes which show magnetic fields was found to increase from f stars to late b stars ( wolff 1968 ) .
the result was recently confirmed by power et al .
( 2007 ) who found an increase of mcp star fractions with stellar mass up to @xmath0 for nearby stars .
a further increase towards higher masses was hidden in the small number of higher - mass stars . while the average fraction of mcp stars among normal stars of the same mass range is generally quoted as being roughly 10% ,
the fraction is less than 2% in the solar neighbourhood of 100 pc radius ( power et al .
why only a fraction of intermediate - mass stars shows strong magnetic fields is one of the key issues in the search for their origin .
the other is that mcp stars form a distinct sub - group with slow rotation among a and b stars .
there are indications of a decrease of magnetic field strength with stellar age for @xmath1 ( kochukhov & bagnulo 2006 ; landstreet et al .
although the increase of stellar radius may cause the reduction at constant magnetic flux , it appears that even the flux is decreasing during the main - sequence life of mcp stars . at first glance
this decay may be attributed to ohmic decay in the radiative envelope of the star .
the radial profile of the magnetic diffusivity @xmath2 throughout the star according to spitzer ( 1956 ) suggests that the decay time is of the order of @xmath3 yr everywhere in the radiation zone . the decay time inferred from landstreet et al .
( 2007 ) is about @xmath4 to @xmath5 yr ( as is @xmath6 in @xmath7/s accidentally ) .
figure [ decay_times ] shows @xmath6 as a function of radius and the corresponding ohmic diffusion time @xmath8 .
the above decay time deduced from observations is also indicated without radial dependence for obvious reasons .
it appears as if the magnetic diffusivity is higher than the purely microscopic value hinting on very mild turbulence even without convection .
another puzzling fact is the decrease of obliquity of the magnetic field axis against the rotation axis with rotation period ( landstreet & mathys 2000 ; bagnulo et al .
mhd processes tend to prefer axisymmetric solutions for fast rotators , if a small degree of differential rotation is present in the star .
this is the very opposite of what is observed . when comparing the rotational time - scale with the diffusive time - scale , even the slowest cp stars have rotation periods many orders of magnitude shorter than @xmath9 and are fast rotators in that sense .
since in view of the time - scales normal a stars and ap stars are not vastly different , we tend to place the magnetic fields first the reduced rotation is a consequence of their presence .
in other words , the reason for mcp stars to be magnetic is not their initially slow rotation .
there is not only the gradual decrease of magnetic field strengths with the time which the @xmath10 stars spent on the main sequence , but also a possible late emergence of surface fields for stars with @xmath11 .
hubrig et al .
( 2000 ) found that these stars show magnetic fields only after they have spent about 30% of their life on the main sequence .
meanwhile , counter - examples were found , and the appearance of fields is certainly not as sharp as it seemed at an earlier research stage . it should be noted that in general all these characteristics are based on quite scattered individual stars .
the scatter is very likely not a result of observing uncertainties .
any physical process leading to the observed features must also allow for a certain bandwidth of results or sufficient input of stochastic behaviour .
in a largely radiative environment , non - turbulent ( laminar ) flows may still be present , namely meridional circulation and differential rotation .
these flows are driven by radiatively ( eddington , 1929 ; sweet , 1950 ; kippenhahn , 1958 ) .
while turbulent convection zones provide very good prospects for dynamo action , one has to search for dynamo solutions in laminar flows , if an mhd dynamo is considered an option for the presence of magnetic fields on cp stars .
solutions for special flow constructions are known for more than 50 years ( e.g. bullard & gellman , 1954 ) and always provide non - axisymmetric solutions according to the theorem of cowling ( 1934 ) .
a comprehensive compilation of spherical flows and their dynamo solutions was given by dudley & james ( 1989 , hereafter dj ) .
the cowling theorem says that an axisymmetric magnetic field can not be sustained by dynamo action .
this is essential for laminar dynamos , while it is unimportant for turbulent dynamos where nonaxisymmetries will always be present in small - scale fields .
the rotation rate is typically normalized in terms of the magnetic reynolds number @xmath12 where @xmath13 is the radius of the sphere and @xmath14 is a typical angular velocity .
dj constructed a very simple flow which does provide a dynamo for fairly low rm .
all these dynamo models are kinematic ; the equation of motion with the corresponding lorentz force is not considered in any of them . only the induction equation @xmath15 with a prescribed @xmath16 .
this is suitable for searches for the onset of dynamo action ; the full nonlinear evolution of the dynamo also requires the momentum equation with a lorentz force .
all the dynamo and stability results obtained for this paper were computed with the mhd code by hollerbach ( 2000 ) .
the original dj flow has an @xmath17 profile decreasing with radius and a meridional circulation which points towards the equator near the surface .
figure [ dudley_james ] shows the critical rm for the onset of dynamo action as a function of the meridional flow speed .
the amplitude of the meridional flow velocity is measured as a fraction of the maximum azimuthal velocity , @xmath18 .
the original dj system for a full sphere is shown as a solid line . in an a star
, we expect a convective core with very strong turbulent diffusion acting like a vacuum hole .
we have modified the flow to obey a central hole and have plotted the critical rm lines for core sizes of @xmath19 and @xmath20 .
the area in which dynamo action is found diminishes and is entirely gone for an inner hole of @xmath21 .
moreover , a meridional flow consistent with a negative @xmath14 gradient would actually be poleward below the surface .
there is no dynamo action for such a flow , but it is retained adding latitudinal differential rotation ( arlt , 2006 ) . for another class of meridional flows without any differential rotation , moss ( 2006 ) did find dynamo solutions , all with critical rm beyond @xmath22 .
since all the solutions found so far consist of rather simple @xmath23 modes , the diversity of magnetic field configurations observed can probably not be covered by such dynamos .
there are essentially two types of instabilities which are relevant for magnetic fields in radiative envelopes .
one is the magnetorotational instability ( mri ; velikhov , 1959 ; balbus & hawley , 1991 ) which relies on a differential rotation decreasing with distance from the axis , and the other class is curent - driven instabilities for which rotation is only a modifier . especially in the beginning of the existence of the radiative envelope , the magnetorotational instability may be of interest for mcp sars .
only small magnetic fields are required for the onset of the instability .
since the energy source for the instability is the differential rotation instead of the magnetic field , perturbations may grow up to considerable strength in fairly short times .
the stability analysis for nearly all magnetic - field configurations shows instability for field strengths which are not too strong ( balbus & hawley , 1998 ) .
how strong is that ?
the instability becomes significantly weaker if the scale of the dominant unstable mode is larger than the object size . in a cp star with a differential rotation of as small as 5%
, the maximum field strength is as large as 20 kg ; it is larger than 100 kg for 30% differential rotataion . even though the mri is a weak - field instability it has a huge range of applicability in a stellar context .
arlt et al .
( 2003 ) have studied such a scenario in a spherical shell .
the setup contained a magnetic field perturbation with both axisymmetric and non - axisymmetric parts , and without symmetry about the equator , in order to allow for a variety of modes to grow .
the rotation rate decreases with axis distance .
the model is simplified in that it employs an incompressible spherical shell with constant density , but allows for a subadiabatic temperature profile as is present in radiative zones ( boussinesq approximation ) .
the equations now read @xmath24 where @xmath25 is the kinematic viscosity , @xmath26 is the pressure , @xmath27 is the thermal expansion coefficient , @xmath28 denotes temperature deviations from a purely conductive background profile @xmath29 , @xmath30 is the gravitational acceleration , and @xmath31 is the thermal diffusivity . the constants @xmath32 and @xmath33 are the magnetic permeability and the density , respectively .
the full three - dimensional and nonlinear evolution of an initial rotation profile and magnetic field is followed .
the mri sets in much faster than the diffusive time - scale , and the fluctuations of @xmath34 and @xmath35 start to redistribute angular momentum within the radiative zone by reynolds stresses ( correlation of velocity fluctuations ) and maxwell stresses ( correlation of magnetic - field fluctuations ) .
the model was modified to include a background density stratification , sacrificing the temperature equation though .
the time - scale for the redistribution of angular momentum remains the same as for the model with a stable temperture gradient .
a snap - shot of a meridional section of the resulting flows is shown in figure [ sps_apdec_d2 ] . at that point
the angular momentum is already entirely redistributed towards a uniform rotation .
realistic magnetic reynolds numbers are not accessible in a time - dependent numerical simulation .
one always has to compute a set of scenarios and extrapolate the results such as magnetic field strength and decay time for the differential rotation to stellar parameters .
the calculations presented here lead to a time - scale between 10 and 100 myr for the decay of any initial differential rotation in mcp stars .
the other class of instabilities is not based on the differential rotation , but on the presence of currents in the initial magnetic field .
nearly all poloidal magnetic fields become unstable when imposing a small non - axisymmetric perturbation ( cf .
braithwaite , 2007 , and references therein ) .
the same holds for purely toroidal magnetic fields which were studied by vandakurov ( 1972 ) and tayler ( 1973 ) .
solid - body rotation has a stabilizing effect on poloidal as well as toroidal magnetic fields .
for this review , the dependence of the stability of a purely poloidal field on the rotation rate was computed .
the normalized rotation rate is again expressed by the magnetic reynolds number .
the stability limit versus rm is plotted in fig .
[ poloidal_field ] .
the hartmann number is the dimensionless magnetic field strength ; it needs to be converted to physical values for specific objects . at high rm the ones in which we are interested in the case of cp stars
the critical magnetic field strength becomes a power function of rm .
we can try to extrapolate the graph to the very high stellar magnetic reynolds numbers of cp stars .
the result scales to a critical field strength of about 10 g for a rotation period of 1 day , and to about 1 g for a 100-day period .
maximum field strengths of up to 100 g were found for purely toroidal fields in the solar radiation zone ( arlt et al . , 2007 ) .
all these limits are much below the observed field strengths .
it is possible construct a magnetic field of unlimited stability .
chandrasekhar & kendall ( 1957 ) reported about an entirely force - free field , and we confirmed its stability numerically . the construction requires the sphere to be embedded in a perfectly conducting medium .
if vacuum is employed , the stability is lost , and surface forces emerge with unkown effect . since the nonlinear simulations of braithwaite & nordlund ( 2006 ) lead to a configuration very similar to the force - free field , further studies are worthwhile in this direction .
the early contraction phase of star formation may host a turbulent dynamo which can provide kg - fields .
once the star becomes radiative , the generated field could become frozen in and visible at the surface when the convective shell has vanished .
the idea is supported by the findings of dikpati et al .
( 2006 ) who showed that the oscillatory dynamo of the solar convection zone can lead to the build - up of stationary fields in the radiative solar interior .
the following numerical experiment probes the applicability of such a mechanism for mcp stars .
a similar approach was followed by kitchatinov et al .
( 2001 ) for solar - type stars .
we start with a fully convective sphere in which the @xmath27-effect is believed to operate as a field generator .
the @xmath27-effect is the description of the generation of poloidal magnetic fields from toroidal ones and vice versa in rotating , stratified turbulence .
as time progresses , the convection zone becomes a shell , and the inner radius of the shell increases gradually
. the core is radiative which is represented numerically by a vanishing @xmath27-effect and a magnetic diffusivity which is 1000 times smaller than the turbulent magnetic diffusivity @xmath36 in the convection zone .
the induction equation for a dynamo with a magnetic diffusivity depending on the location in the star is @xmath37 where @xmath16 can contain all large - scale velocities such as differential rotation and meridional circulation .
the @xmath27 effect is distributed such that it is proportional to @xmath38 in the convective shell as well as in the inner convective core forming near the end of the simulation , where @xmath39 is the colatitude .
we also introduced an anisotropy in the @xmath27-tensor as expected from rotating turbulence , such that the @xmath27-component parallel to the rotation axis vanishes ( rdiger & kitchatinov , 1993 ) .
the last term in eq . ( [ induction_alpha ] ) is due to the turbulence gradient and can be described as turbulent pumping since it has the effect of a velocity ( zeldovich , 1957 ; ruzmaikin & vainshtein , 1978 ) .
the term is often neglected in computations of turbulent dynamos with functions @xmath40 , but plays a significant role in confining magnetic fields in radiative interiors . the temporal evolution in fig . [ alpha2omega_evol_ap ] shows a short phase with dominance of the non - axisymmetric mode .
this agrees with earlier findings that fully convective shells may provide @xmath23 dynamo modes .
when the radiative core becomes larger than @xmath41 , the @xmath42 dominates and does so for the remaining time of the simulation .
the storage of magnetic flux in the radiative interior appears to be chiefly axisymmetric , unfortunately .
an initial convective dynamo may thus not provide the observed fields directly .
the onset of a core dynamo does produce strong internal fields , but these weaken severely towards larger distances from the center .
this was also found from direct numerical simulations by brun et al .
an issue we have not addressed in this review so far is the survival of large scale magnetic fields from before the collapse of the star .
although we concentrated on dynamos and instabilities , a few words may be necessary for a comprehensive picture .
a fully convective phase suggested to take place for stars with masses below @xmath43 ( palla & stahler , 1993 ) could destroy these primordial fields in a rather short time of a few thousand years due to turbulent diffusion .
moss ( 2003 ) argued , however , that the exceptional strength of the fields in mcp star progenitors reduces turbulent diffusion and allows the survival of sufficient flux to be obervable after the star has reached the zero - age main - sequence .
the exact behaviour of the diffusion reduction known as @xmath6-quenching in mean - field electrodynamics is not known though .
even if it is not largely convective , the proto - stellar object will also be subject to differential rotation .
torques exerted by accreted material , magnetic coupling to the accretion disk and magnetic winds very likely cause differential rotation in the star ( stpie , 2000 ) .
primordial fields of arbitrary direction will excite the magnetorotational instability as long as the magnetic fields are weaker than @xmath44 kg which is probably true .
the instability is actually a good candicate for the diversity if field configurations afterwards ; it also allows for amplification of the field , with the energy being taken from the shear .
a compilation of a few possible mechanism in cp star magnetism with their pros and contras are listed in table [ summary ] .
when searching for mechanisms explaining considerable magnetic fields in radiative envelopes , one always returns to the question why only a small fraction of a stars shows such fields .
the effect of different star formation environments on the existence of strong fields on main - sequence stars now needs to be investigated .
bridging collapse and pre - main - sequence evolution is not a trivial task though . | a variety of magnetohydrodynamic mechanisms that may play a role in magnetic , chemically peculiar ( mcp ) stars is reviewed . these involve dynamo mechanisms in laminar flows as well as turbulent environments , and magnetic instabilities of poloidal and toroidal fields as well as combinations of the two .
while the proto - stellar phase makes the survival of primordial fields difficult , the variety of magnetic field configurations on mcp stars may be an indication for that they are instability remnants , but there is no process which is clearly superior in explaining the strong fields . |
low - dimensional disordered excitonic systems have generated renewed interest recently .
for example , the possibility of using optically active polymers in various photonic applications such as flat panel displays and light emitting diodes has led to increased interest in their electronic and optical properties @xcite .
we study here numerically the dynamics of frenkel excitons near the band center in a system with nearest - neighbor off - diagonal random interactions .
this study is , we believe , the first to treat correlated disorder in the off - diagonal dipolar interactions in an exciton system which focuses on the behavior near band center @xmath1 .
however , an analysis of a similar problem in the correlated electronic system has appeared @xcite but the authors considered correlations those were different from ours , and the interactions were not of the dipolar type @xcite .
many efforts have also been spent on the same off - diagonal tight - binding system using supersymmetric methods ( susy ) @xcite where the interactions are formally similar to our dipolar interactions but the authors eventually worked with continuous variables whereas ours are discrete . later , correlated disorder , in particular the exponential type , was included for the same tight binding model using the susy @xcite .
our results can thus be compared with what has been developed for the susy methods .
this study is important because most realistic random disordered systems have nonlocal correlations .
the nature of the excitations at the band center of random off - diagonal exciton models has been discussed for a long time , and it has been found that the spectrum has a singularity of the dyson type @xcite in which spectral properties in the vicinity of the band center are functions of @xmath2 where @xmath3 is the energy relative to the center of the band . the question of the band center localization has also been argued at length , and it was claimed that the band center mode may be weakly localized because of the strong fluctuations @xcite .
ziman @xcite analyzing various uncorrelated distributions of the random coupling found the low - energy behavior of the integrated density of states ( @xmath4 ) @xmath5 and the inverse localization length ( @xmath6 ) @xmath7 in which @xmath3 is the energy and the logarithmic variance @xcite @xmath8 } $ ] is the amplitude factor , @xmath9 being the off - diagonal coupling specified below .
for definitions of the @xmath6 and the @xmath4 see ziman s work @xcite particularly his discussion after the equation of the motion .
notice that the averages above are independent of the lattice site @xmath10 since uncorrelated random variables are involved ; otherwise they are site dependent .
concerning the shape of the @xmath4 and the @xmath6 , according to ziman s analysis @xcite , they should depend on disorder only through a multiplicative constant , the variance in the distribution of @xmath11 . as a result ,
the spectra near @xmath12 have the same form for all sets of independent , identically distributed random variables @xmath9 . in this context , such a property is referred to as scaling . when there is scaling , the data can be collapsed onto a single curve by dividing through by a factor that depends on the properties of the disorder but
is independent of the energy .
although the above mentioned results were developed for uncorrelated disorder , our analysis shows that qualitatively similar scaling behavior for limited range of concentrations and energy can be obtained for a system with exponential correlations but with a modified amplitude factor @xmath13 .
moreover , for a certain range of correlation lengths a higher order expansion of the dyson singularity can reproduce the numerical data better than just a single term . a disordered 1d exciton model with nearest - neighbor interactions has the following equation of motion @xmath14 where @xmath15 denotes the usual exciton creation operator and @xmath16 are the couplings .
this is equivalent to the xy model in a strong applied field in the @xmath17-direction ( @xmath18 ) . in our numerics we calculated iteratively
the lyapunov exponent @xcite @xmath19 ( see appendix ) , where @xmath20 is the number of sites @xcite , whose real part is related to the @xmath6 while the imaginary is related to the @xmath4 .
we consider couplings that have correlated or uncorrelated ( for comparison ) disorder and are of a dipolar type . since the displacements of the atoms in quasi - one dimensional systems are likely to be small compared to the lattice constants , it is usually enough to consider the first order term in the expansion of @xmath9 in terms of the displacements @xcite .
for the first order expanded coupling we have @xmath21 with unit lattice constant where @xmath22 is the strength of the disorder and @xmath23 is a random sign variable that is exponentially correlated @xmath24 where @xmath25 is the correlation length as a function of `` wrong sign '' concentration @xmath26 ( see below ) .
thus the dipolar coupling @xmath27 has a non fluctuating part and a fluctuating part as in dirac fermions with random - varying mass @xcite .
the generation of random binary sequences with correlation was deviced by implementing the convolution method of constructing a random squence .
such squence can be used in various applications as in designing low - dimensional devices with desired properties @xcite .
a detailed discussion on this topic recently appeared @xcite .
however , a more efficient and easily handled correlated sequence , related to the spin glass chain problem @xcite , can be obtained through uncorrelated random numbers as outlined below .
the correlated random number at site @xmath10 follows the relation @xmath28 with the distribution ( of @xmath29 @xmath30 within @xmath31 $ ] ) @xmath32 the @xmath29 are uncorrelated between different sites and @xmath26 is the concentration of `` wrong signs '' such that @xmath33 and @xmath34 mean no disorder .
clearly @xmath35 is exponentially correlated @xmath36 where one can define a correlation length @xcite @xmath37 . for the numerical computation
uncorrelated @xmath29 are obtained from the uniformly distributed random number generator sequence . for the uncorrelated results we set @xmath38 at the particular concentration @xmath39 where both correlated and uncorrelated cases produce identical results since @xmath40 .
it should be emphasized that while the model for uncorrelated disorder and the model for correlated disorder involve the parameter @xmath26 , in the latter model @xmath26 determines both the fraction of ` wrong sign ' bonds and the correlation length .
the two models ` intersect ' at @xmath41 where there are equal numbers of @xmath42 interactions and the correlation length is @xmath43 .
we have carried out extensive numerical studies to verify the equation for the correlation length given above .
we note here that unlike the off - diagonal case above , many works have recently appeared on the diagonal tight binding model with correlated disorder dealing with various aspects of the problem @xcite .
it was revealed that even short range correlated disorder at the band edge caused the anomalies in the spectrum @xcite .
frenkel excitons in off - diagonal and diagonal models have markedly different spectra , i.e. , the former has a dysonian singularity @xcite at the band center while the latter has a power - law singularity at the band edge .
hence the short - range correlations are expected to influence off - diagonal spectra more than in the diagonal case .
the distribution and localization of the exciton modes are characterized by the lyapunov exponent . in the case of one - dimensional arrays with nearest - neighbor interactions ,
the lyapunov exponent can be determined by making use of mode - counting techniques @xcite .
recently we studied @xcite similar excitons with orientationally disordered couplings in conjugated polymers and calculated the full spectrum and optical line shapes .
the goal here is to look at the effects of exponentially correlated disorder on the behavior of the @xmath4 and @xmath6 near the center of the band ( @xmath1 ) .
we have produced three sets of data for each case varying the correlation length i.e. , c ) and the strength of the disorder , @xmath44 with @xmath45 .
( 15,5 ) = 7.5 cm figure [ fig : ids0 ] displays the results of the @xmath6 and @xmath4 for exponentially correlated distributions of the couplings . for fixed @xmath22 , we divided the data produced for a particular concentration @xmath26 by a scaling factor @xmath46 ( see next section and the appendix ) .
the data for selected concentrations are shown in fig .
[ fig : ids0 ] for the indicated values of @xmath22 .
we expect first that for a range of concentrations the data should collapse to a dysonian singularity for fixed @xmath22
. however , fig .
[ fig : ids0 ] reveals that as @xmath26 increases , the @xmath6 and the @xmath4 take values that differ increasingly from predictions based on ziman s analysis @xcite and a slow approach to the band center is observed . in this limit
, the data clearly show qualitative scaling behavior but deviations with varying magnitude exist as @xmath47 and @xmath48 where the correlation length is infinite .
notice also that the deviations are stronger for the @xmath4 than for the @xmath6 . clearly as
energy decreases , the scaling gets better ; however , the deviations become worse as one gets closer to @xmath49 . to see whether longer chains and/or lower energies improve the situation , we checked further and found that the longer chain did not improve the agreement appreciably , but as the energy is lowered , the results follow the scaling behavior more closely up to a certain limiting @xmath26 . similar disagreement between the data and the dysonian ( with a single term )
singularity were encountered in the susy @xcite off - diagonal tight binding case for continuous exponential correlations .
the analytical investigations @xcite revealed that the spectra for the exponential correlation can be expanded for the @xmath4 as @xmath50 and for the inverse localization length ( @xmath6 ) @xmath51 where @xmath52 are given as of functions of the correlation length multiplied by the strength @xcite of the exponential correlations .
the equations ( [ idos])([ill ] ) are only the first order term in the expansion and @xmath53 and @xmath54 should correspond to correlated amplitude factor @xmath55 .
figure [ fig : idsf ] clearly shows the improved results obtained using the eqs .
( [ idosc])([illc ] ) for the selected concentrations where the worst deviations followed from the first order result : @xmath56 .
unfortunately , for @xmath57 and @xmath58 , this type of expansion appears to fail ; however , other forms of fitting with fractional powers can give better results such as tried in the susy case @xcite .
we will not here discus significance of this fractional fitting since there does nt seem to be any analytic theory that predicts it . [
fig : idsf ] ( 15,5 ) = 7.5 cm
to shed some light on the scaling behavior for the correlated case , we first compute the binary uncorrelated case . a simple calculation of the logarithmic variance , the amplitude of the spectrum , for the uncorrelated case given in eqs . ( [ idos])([ill ] ) yields the result @xmath59 the amplitude factor for the uncorrelated disorder clearly displays a scaling behavior in concentration @xmath26 as well as in @xmath22
we can identify the scaling functions for the concentration @xmath60 and the strength of the deviation @xmath61 .
the top inset in fig .
[ fig : scl ] displays the uncorrelated data vs concentrations .
the three lines are the calculated amplitudes @xmath62 presented in eq .
( [ uvar ] ) for @xmath63 .
as expected , the theory for the uncorrelated case works well the amplitude factors in fig .
[ fig : scl ] ( large panel ) are found from the best fit to the dysonian singularity .
they are obtained matching the smallest value of the energy rather than the value obtained by fitting over a range of energies . in this way
, we will be more likely to get a value characterizing the asymptotic region . the calculated amplitude scaling function is also shown in fig .
[ fig : scl ] for the correlated case @xmath64 for the three specified @xmath22 values above . in the appendix
we will argue why this formula holds .
we notice immediately that the concentration dependence has a different form @xmath65 as compare to @xmath60 of the uncorrelated case while the scaling on @xmath22 is the same for both cases .
this asymptotic fit can further be checked by plotting the @xmath53 and @xmath54 in the expansion above since they correspond the coefficient of the dysonian singularity .
the bottom inset displays the @xmath53 and @xmath54 obtained by the best fit as compare to the theoretical line @xmath55 ( for @xmath67 ) .
both plots clearly support the asymptotic amplitude scaling .
[ fig : scl ] ( 15,5 ) = 7.5 cm although the results obtained by asymptotic fit are clearly reproduced by the theoretical amplitude factor , there are striking differences between correlated and uncorrelated cases .
what is particularly interesting and not expected is that the results for correlated disorder is not symmetric about @xmath39 like those of the uncorrelated case .
also , from our scaling analysis the @xmath6 coefficient becomes large as @xmath47 , but vanishes as @xmath48 .
this is surprising since the modes for @xmath34 are extended just like the modes for @xmath33 .
the divergence of the amplitude factor as @xmath68 is not inconsistent with the results for @xmath34 since the energy interval over which scaling holds shrinks to zero in the same limit .
this can be seen from eqs .
( [ idos])([ill ] ) by setting either the @xmath4 or the @xmath6 to a constant value and solving for the cut - off energy as as a function of @xmath26 . as shown in the appendix ,
the amplitude factors perform a correlated random walk in parameter space . for our case ,
the square of the length of the random walk is proportional to @xmath69 not @xmath20 alone as in the uncorrelated case . that is the source of the marked difference between the two cases . as the concentration approaches @xmath70 ,
divergent behavior occurs .
however , we note that we can not take c arbitrarily close to @xmath70 ; beyond a certain point , the random walk sequence can not be constructed @xcite .
we have investigated the dynamics of 1d frenkel exciton systems with correlated off - diagonal disorder . we have used a negative eigenvalue counting technique @xcite which provides a simple and physically transparent analysis of the @xmath4 and the @xmath6 .
we investigated the question of scaling when there was correlated off - diagonal disorder such as might occur , for example , when the off - diagonal interaction depended on the distance between two ions and thus would be affected by small displacements of the ions from their equilibrium positions @xcite .
our numerical data indicate that in the correlated case the scaling behavior found by ziman @xcite is followed only for a limited range of @xmath26 and very small @xmath3 .
we compared our results with the scaling predictions of ziman @xcite and with the similar tight binding correlated electronic case @xcite .
we computed the asymptotic amplitude factors as a function of concentration and found that they are equal to the uncorrelated variance at @xmath39 multiplied by the factor @xmath71 ( see appendix ) .
this factor @xcite has played a very interesting scaling role in number of unrelated problems .
we found that with the new variance for correlated distributions rather unexpected behavior was obtained as compare to that of the uncorrelated case .
in particular , the asymptotic amplitude factor is not symmetric about @xmath39 while the @xmath6 diverges for concentrations approaching @xmath70 in contrast to what is observed in the uncorrelated case .
this amplitude factor can explain only the first order expansion in a dyson - type singularity as supported by the susy model @xcite . as shown in the appendix , if the random walk observation of eggarter _ _ et__@xmath72_al .
is _ implemented , with correlation , the obtained amplitude factor can reproduce the data within a shrinking energy range as @xmath73 as @xmath26 is increased .
finally , we should emphasize that our work is a numerical study supplemented by the approximate theoretical analysis outlined in the appendix . as mentioned previously
, the surprising result is the discovery that near the center of the exciton band , a model with correlated disorder showed asymptotic behavior that is similar to the behavior of systems without correlations , the only difference being in the amplitude factor appearing in the limiting expressions for the idos and ill . like those of essentially all numerical studies , our results are approximate .
we hope that the availability of the numerical findings together with our analytical results will stimulate rigorous analyses of the model that may shed light on the origin of the similarity .
this work is partially sponsored by the scientific and technical research council of turkey ( tubitak ) and ege university research grant ( bap ) .
in this section we present the amplitude scaling related to the first order expansion coefficients @xmath53 and @xmath54 in eqs .
( [ idosc])([illc ] ) . similar
scaling but in different cases has appeared before @xcite .
however the off - diagonal problem here is more involved than those cases .
first the lyapunov exponent at the band center @xmath1 is calculated using eq .
( [ tb ] ) that can be rearranged @xcite the bipartite nature ( the chiral symmetry ) of the eq .
( [ tb ] ) is responsible for the logarithmic variance as rigorously shown in @xcite and eventually the dyson- type singularity @xcite since @xmath75 executes a random walk with a step @xmath76 . for the uncorrelated case the averages yield @xmath77 and @xmath78 $ ] .
the lyapunov exponent @xcite takes this form @xmath79 .
the imaginary part here @xmath80 and the real part is related to the @xmath6 . to make our point , we can attack the problem with a different angle . the average of the summation @xmath81 for both correlated and uncorrelated cases ; whereas , the average of the squared summation @xmath82 .
it further reduces to @xmath83 since the second sum is zero for the independent random numbers but non zero for the correlated random numbers . for the correlated disorder
the summation is not easy .
the complication arises when correlation present since @xmath84 is not equal to its ensemble average rather it dependents on @xmath10 . but this complication can be avoided in a following way .
let us calculate the taylor expansion of @xmath85 which reads @xmath86 the first sum ( with odd powers of a ) is just the expansion of @xmath90 and let the second sum ( with even powers ) be @xmath91 which will be eliminated when it is inserted in @xmath92 .
the same expansion can be used to calculate the uncorrelated case as well ; the step takes the form @xmath93 ( recall that number of steps @xmath94 ) . using eq .
( [ dist ] ) , the variance is given by @xmath95 so that the @xmath6 becomes @xmath96 .
for the correlated case , we have @xmath97 . taking the square
, we get three terms @xmath98 and twice @xmath99 , hence the summation results in @xmath100 where inverse of this factor is encountered in previous works @xcite
. the correlated random walk results in @xmath101 and the @xmath102 . in contrast , the @xmath103 value at the band center is not affected by the correlations since the distribution in eigenvalues is symmetric about the mid - point .
the calculation above holds at the band center ; however eggarter
_ _ et__@xmath72 _ al .
_ , @xcite have shown that for a certain range of @xmath3 close to the band center , the random walk behavior still held and the @xmath6 and @xmath103 can be represented as in eqs .
( [ idosc])([illc ] ) where the amplitude factor is the same as calculated at the band center . according to the frenkel exciton model ,
the transfer integral usual dipole - dipole interaction has the inverse cube of the distance between the dipoles .
offdiagonal transfer integral disorder is due to the distribution of bond length . assuming the transition dipoles , the intersite transfer dipole - dipole interaction is given as @xmath104 where @xmath105 is the positional deviation from their equilibrium position @xmath106 of transition dipoles .
the nearest distance between the atoms is @xmath107 and @xmath10 is the site index .
since the interaction decreases rapidly over the @xmath10 to @xmath108 distance , regarding nearest - neighbor couplings is sufficient as a first approximation .
assuming unit neighbor distance the tranfer integral reduces to @xmath109 where @xmath110 is the strength and @xmath111 is the random sign of the deviation from the regular lattice point .
notice that @xmath112 is used in the text as @xmath113 .
the deviations from the equilibrium positions introduces a correlation .
if @xmath110 is small , one can expand the coupling as @xmath114 . since b is a free parameter , one can replace 3b with a without loss of generality .
note that using the approximate form in place of the cubic form does not make an essential change in the spectra apart from a numerical factor since the spectra are related to averages of the logarithm , @xmath115 and our numerical study confirmed this . | we report the amplitude scaling behavior of frenkel exciton chains with nearest - neighbor correlated off - diagonal random interactions .
the band center spectrum and its localization properties are investigated through the integrated density of states and the inverse localization length .
the correlated random interactions are produced through a binary sequence similar to the interactions in spin glass chains .
we produced sets of data with different interaction strength and `` wrong '' sign concentrations that collapsed after scaling to the predictions of a theory developed earlier for dirac fermions with random - varying mass .
we found good agreement as the energy approaches the band center for a wide range of concentrations .
we have also established the concentration dependence of the lowest order expansion coefficient of the scaling amplitudes for the correlated case .
the correlation causes unusual behavior of the spectra , i.e. , deviations from the dyson - type singularity .
+ exciton , one dimension , correlated off - diagonal disorder , amplitude scaling 78.20.bh , 78.40.pg,*78.67.@xmath0n * |
recently , the pamela collaboration has observed an anomalous positron abundance in cosmic radiation @xcite .
the positron over electron fraction turns over and appears to rise at energies from 10 gev to 100 gev .
however , from the same detector , no obvious antiproton excess is seen for the same energy range @xcite .
many suggestions have been made to explain the positron excess at pamela . among different approaches ,
dark matter annihilation is especially interesting and could imply future signals in dark matter direct detection experiments and/or at the large hadron collider ( lhc ) . in recent model - independent studies ,
dark matter particles are required both to annihilate dominantly to leptons and to have a much larger annihilation rate in the galactic halo than would be implied by a traditional thermal relic estimate @xcite @xcite .
the lack of antiproton excess in the pamela experiment @xcite can be explained if dark matter particles first annihilate into some intermediate particles , which are so light that their decays to hadrons are kinematically forbidden @xcite , or couple dominantly to standard model ( sm ) leptons @xcite . to explain the large dark matter annihilation cross section in the galactic halo while remaining consistent with a thermal dark matter relic abundance
, one can introduce an attractive force between two dark matter particles and use the sommerfeld enhancement to boost the annihilation cross section in the galactic halo @xcite .
this scenario suggests interesting signatures @xcite and can be tested at the lhc @xcite .
a second approach is to consider non - thermal relics .
other long - lived particles can decay to the dark matter particles and increase the dark matter relic abundance in the galactic halo @xcite .
instead of constructing a dark matter model _ ad hoc _ to explain the cosmic ray observations , in this paper we ask whether there is an existing well - motivated model that naturally contains the necessary ingredients to explain the positron excess of pamela .
the most developed framework to address the naturalness problem of the sm is the minimal supersymmetric standard model ( mssm ) , which provides the lightest superpartner ( lsp ) protected by r - parity as the dark matter candidate . in general , the annihilation products in the mssm contain not only leptons but also a large fraction of hadrons .
this makes it difficult for the mssm to explain the positron excess at pamela , and at the same time be consistent with the antiproton spectrum at pamela @xcite . from the theoretic side ,
the mssm suffers the @xmath5-problem , which can be solved elegantly by introducing a new gauge singlet chiral supermultiplet , as proposed in the next - to - minimal supersymmetric standard model @xcite .
a recent exploration of the nmssm parameter space shows that the lightest @xmath0-odd particle @xmath2 ( mainly from the singlet component ) can be naturally lighter than @xmath6 and mainly decay into two @xmath7 s @xcite .
one should notice that the mass of @xmath2 in the nmssm is protected by the @xmath8 symmetry and can even be lighter than @xmath9 gev if the soft terms associated with the singlet are small .
in this case the @xmath2 will decay mainly into two muons for a mass of a few hundred mev .
therefore , if the dark matter candidate in the nmssm can annihilate mostly into @xmath2 s , we can have a leptonic final state in the annihilation products simply from kinematics .
the nmssm provides the necessary ingredients to make this happen . notice that the dark matter candidate in the nmssm is the lsp neutralino ( for existing studies for light neutralino dark matter in the nmssm , see @xcite ) .
if its mass happens to be around half that of the heavier @xmath0-odd scalar @xmath4 mass , a large dark matter annihilation cross section is obtained through the @xmath3-channel resonance effects with @xmath4 . to have leptons dominant in the final state
, we should have a large branching ratio of @xmath4 to @xmath2 plus @xmath1 .
this will naturally happen , provided that the dark matter lsp mass is less than the top quark mass , and thus that the decay of @xmath4 to @xmath10 is kinematically forbidden . since the @xmath0-odd scalar coupling to other fermions is proportional to their yukawa couplings , the final state can dominantly be @xmath11 with the former decaying to leptons .
the kinematics helps us to obtain hard leptons in the final state of the dark matter annihilations . in section [ sec :
model ] , we develop the notation by deriving the spectrum and interactions in the nmssm , and show two sets of representative model points allowed by current experimental constraints .
we calculate the dark matter annihilation cross section in section [ sec : annihilation ] and positron excesses from neutralino annihilation in section [ sec : positrons ] . in section [ sec : pamela - proton ]
, we consider the constraints on the model parameter space from the pamela antiproton spectrum .
we discuss the gamma - ray spectrum in section [ sec : gammaray ] and point out a discrepancy with recent preliminary fermi lat results and comment on a possible resolution . in section [ sec : constraints ] we demonstrate the consistency of our model points with direct constraints from lep , tevatron , cleo , b - factories and the magnetic moment of the muon .
finally , we discuss dark matter direct detection and conclude in section [ sec : conclusions ] .
to describe the nmssm model , we follow the notation in the ref . @xcite .
the superpotential in the nmssm is w=+^3 , and the soft supersymmetry - breaking terms are v = a_sh_uh_d+a_s^3+h.c .. here , hatted capital letters denote superfields , and unhatted capital letters the corresponding scalar components .
the minimization of the scalar potential determines their vacuum expectation values ( vevs ) : @xmath12 , @xmath13 and @xmath14 .
the electroweak scale @xmath15 gev .
since @xmath16 , there are four new parameters in the nmssm , which we take to be real : @xmath17 , @xmath18 , @xmath19 and @xmath20 . with sign conventions for the fields , @xmath17 and @xmath21
are positive , while @xmath18 , @xmath19 , @xmath20 and @xmath22 can have either sign .
there exists a @xmath23 symmetry for the nmssm , which is spontaneously broken and induces a domain wall problem .
one can introduce higher dimension operators to explicitly break this discrete symmetry and perhaps circumvent this problem @xcite .
since those @xmath23 breaking operators have small effects on the analysis performed in this paper , we will neglect them from now on .
there are five neutralinos in the nmssm : the @xmath24 gaugino @xmath25 , the neutral @xmath26 gaugino @xmath27 , the higgsinos @xmath28 and @xmath29 and the singlino @xmath30 . in the basis @xmath31
, we have the neutralino mass matrix = -(^0)^tm_0(^0)+h.c . , where _ 0= ( ccccc m_1 & 0 & & - & 0 + 0 & m_2 & - & & 0 + & - & 0 & - & - h_d + - & & - & 0 & - h_u + 0 & 0 & - h_d & - h_u & 2s ) . to obtain the needed dark matter annihilation rate , we consider the case that the lsp , the lightest neutralino , is mainly made of the bino with mixings with higgsinos . for @xmath32 , a moderate @xmath33 and @xmath34 ,
the lightest neutralino is approximately & & -i_1-(-)^0_u + & & -(+)^0_d , with _ 1&= & , [ eq : alpha1 ] + _ 2&= & , [ eq : alpha2 ] and with its mass approximated as m_=m_1 + .
hereafter , we use a simple notation @xmath35 to replace the usual notation @xmath36 for the lightest neutralino .
the charged higgs @xmath37 has a mass m^2_h^=s(a_+s)+(-^2)v^2 . for @xmath19 and @xmath17 of order of unity ,
the mass of @xmath38 is generically @xmath39 .
expanding around the higgs fields vevs , the neutral scalar fields are defined as h^0_u&=&h_u+ , + h^0_d&=&h_d+ , + s&=&s+ . for the three @xmath0-even neutral states , we can diagonalize their @xmath40 mass matrix by an orthogonal matrix @xmath41 to obtain the mass eigenstates ( ordered in mass ) : @xmath42 , with masses denoted by @xmath43 . for @xmath44 ,
we write the lightest @xmath0-even higgs as h_1&&h_ur+h_dr + & & -s_r,[eq : h1 ] with & & -- , + _
s&=&-(1-)+o(v^3/s^3 ) .
the singlet component of @xmath1 is small and suppressed by @xmath45 .
there are three @xmath0-odd pseudoscalar fields , one of which is a massless goldstone mode eaten by the @xmath46 boson . dropping the goldstone mode , the remaining @xmath47 mass matrix in the @xmath48 basis with @xmath49 , is _ odd= ( cc s(a_+s ) & ( a_-2s ) + ( a_-2s ) & 4h_uh_d+a_-3ka_s ) . we introduce a mixing angle @xmath50 to diagonalize the above matrix : = - , ^2_a= , and arrive at the physical @xmath0-odd states @xmath51 ( ordered in mass ) a_1&= & ( h_ui+h_di)+s_i , + a_2&=&- ( h_ui+h_di)+s_i . + the lightest @xmath0-odd particle @xmath2 is mainly composed of the singlet field when @xmath52 , and is a doublet field otherwise .
since we are looking for a very light scalar , we observe from the determinant of @xmath53 that this occurs if @xmath20 and @xmath18 are small .
this is technically natural since @xmath54 is a symmetry - enhancing limit .
the heavier @xmath0-odd scalar mass is approximately m^2_a_2&= & + & & + . with the condition @xmath55 . for small @xmath18 and @xmath20 , we have the following approximate formulae : m^2_a_1 & & , [ eq : ma1 ] + m^2_a_2&&+2v^2 , [ eq : ma2 ] and & & , + ^2_a & & , for @xmath56 . since we are interested in dark matter annihilation through @xmath3-channel @xmath4 exchange , we list the relevant vertices associated with @xmath0-odd scalars in this section . for @xmath57 ,
the couplings of @xmath0-odd scalars to fermions are & & a_2t_lt^c_r : -i , + & & a_2b_lb^c_r : i , + & & a_1t_lt^c_r : -i , + & & a_1b_lb^c_r : i. the couplings of @xmath4 and @xmath2 to the lightest neutralino are ig_a_2&&ig_1(- ) , + [ eq : a2chichi ] + ig_a_1&&-ig_1(- ) , + [ eq : a1chichi ] with @xmath58 and @xmath59 defined in eqs .
( [ eq : alpha1][eq : alpha2 ] ) .
therefore , the coupling of @xmath60 increases as one increases the higgsino components of the lsp .
their couplings to gauge bosons are & & a_2(p)h^+(p^)w^- _ : -(p - p^ ) _ , + & & a_1(p)h^+(p^)w^- _ : ( p - p^ ) _ , + & & a_2(p)h_1(p^)z _ : -(p^-p ) _ - ( p^- p ) _ , [ eq : a2h1z ] + & & a_1(p)h_1(p^)z _ : ( p^-p ) _ ( p^-p)_. [ eq : a1h1z ] the couplings among @xmath4 , @xmath2 and @xmath1 depend on the diagonalization of the @xmath0-even scalar mass matrix , which may have significant one - loop contributions . for simplicity ,
we use the tree - level results in eq .
( [ eq : h1 ] ) to obtain analytic formulae .
we arrive at the following dimensional couplings to the leading power in @xmath45 @xcite a_1a_1h_1 : & & w_a_1a_1h_1= o(v^3/s^2 ) , + a_2a_1h_1 : & & w_a_2a_1h_1=-+o(^2v^2/s ) . here
the approximation is valid for small values of @xmath18 and @xmath20 . we also need the main decay channels of @xmath1 and @xmath4 . for the @xmath0-even particle @xmath1 with a mass below @xmath61
, it mainly decays into 2 @xmath2 s or 2 @xmath62 s with the decay widths calculated as following : ( h_12a_1)&= & o(v^6/s^4 ) , + ( h_1b+|b ) & & ( ) ^2 .
thus for @xmath63 the decay @xmath64 is suppressed in favor of @xmath65 . note this suppression
is directly connected to the small values of @xmath18 and @xmath20 . for large values of @xmath18 and @xmath20 , the @xmath64 decay would be the dominant one @xcite . similarly for @xmath4 ,
if its mass is below twice the top quark mass , the leading two decay channels are ( a_2h_1a_1 ) & & m_a_2 , + ( a_2b+|b ) & & ( ) ^2 .
the bosonic decay channel can be dominant for modest values for @xmath19 and @xmath17 .
however , if the @xmath4 mass exceeds twice the top quark mass , the decay channel into @xmath10 opens : ( a_2t+|t ) & & ( ) ^2 + & & , and this would become the dominant decay for @xmath4 assuming @xmath66 .
in this section we find nmssm model points that can provide the neutralino as a dm candidate to explain pamela .
there are two relevant possibilities depending on the mass of @xmath2 : 1 .
the mass of the lightest @xmath0-odd particle @xmath2 is in the range @xmath67 .
the mass of the lightest @xmath0-odd particle @xmath2 is in the range @xmath68 . in the first case
@xmath2 mainly decays to two muons because decays to mesons are kinematically suppressed .
because @xmath2 is @xmath0-odd , decays to two pions are forbidden due to cp symmetry , while decays to three pions are suppressed by the three - body phase space . in the following , we will approximate @xmath69 as 100% for case 1 . in the second case , because the couplings of @xmath2 to fermions are proportional to fermion masses , we anticipate that @xmath2 decays mainly to 2 @xmath7 s .
we do not consider the case that @xmath2 mainly decays into two electrons because of stringent constraints from the beam - dump experiment at cern @xcite .
to find the interesting parts of the parameter space , we use the program nmhdecay @xcite for numerical checks .
there are many experimental constraints considered in nmhdecay such as various higgs searches at lep , @xmath70 and @xmath71 .
the model points presented in this paper pass all of the constraints embedded in nmhdecay .
furthermore , we will discuss updated constraints on @xmath72 from babar and searches for dimuon resonances at lep and tevatron in section [ sec : constraints ] .
1.0 @ cccccccc @xmath73 & @xmath17 & @xmath19 & @xmath18 & @xmath20 & @xmath22 & @xmath74 & @xmath75 + 3.1 & 0.24 & 0.194 & -0.05 & -0.273 & -190 & 178.5 & 200 + @xmath76 & @xmath77 & @xmath78 & @xmath79 & @xmath80 & @xmath81 & @xmath82 & @xmath83 + 161.8 & 0.81 & 320.1 & 114.7 & 297.6 & 325.4 & 175.6 & 0.22 + & & & & + & & & & & + & & & & + & + for both cases , in order to isolate the dark matter discussion , we choose the less relevant soft terms to be heavy .
for example , we choose 500 gev soft masses for sleptons , 1 tev for squarks , 1 tev for gluino and @xmath84 tev for all the @xmath85-terms in the quark and lepton sectors .
we have used the updated top quark mass @xmath86 gev @xcite , which has a significant correlation to the higgs boson mass for small @xmath73 as considered in this paper . for the muon - favored case ,
we choose values for other parameters in the nmssm as in table [ tab : case1para ] , which also shows the relevant spectrum and branching ratios of the light scalars .
the default lower limit on the @xmath2 mass is 1 gev in nmhdecay .
one has to change the code file named mhiggs.f to obtain an @xmath2 mass below 1 gev . as can be seen from table [ tab : case1para ] , if @xmath4 can be produced from @xmath87 annihilation through the resonance effect , the final products of the dm annihilation mainly contain @xmath88 .
the @xmath2 decays into two muons to provide the positrons needed to explain the excess at pamela .
the dark matter mass is mainly controlled by the parameter @xmath74 in this model , which is chosen to have the lightest neutralino mass below the top quark mass .
otherwise , @xmath4 will decay into @xmath10 with a significant branching ratio , and the neutralino annihilation produces a limited amount of positrons and lots of hadrons , the hadrons being disfavored by the null antiproton excess at pamela .
it is intriguing that a combination of nmssm and pamela results forces us to have a dark matter mass below around 170 gev .
it is technically natural to have @xmath77 below 1 gev for tiny values of @xmath18 and @xmath20 as reported here , since the @xmath8 symmetry protects its mass ( one can also use the @xmath89 symmetry to obtain a light pseudoscalar , see @xcite for example ) . notice that the branching ratio of @xmath90 is below 1.2% to satisfy the current null results of searches of @xmath2 in the channel @xmath91 at d0 ( see section [ sec : tevatron ] ) . for the tau - favored case , we list the values of model parameters , spectrum of particles and interesting branching ratios of light scalars in table [ tab : case2para ] .
the current direct searches only impose mild constraints on the model parameters .
therefore , we choose one representative point in the parameter space to have @xmath92 as the main decay channel of @xmath1 , and hence to have six @xmath7 s in the final state of dark matter annihilations .
1.0 @ c c c c c c c c @xmath73 & @xmath17 & @xmath19 & @xmath18 & @xmath20 & @xmath22 & @xmath74 & @xmath75 + 2.0 & 0.519 & 0.458 & -14.83 & -3.6 & -200.0 & 162.25 & 1000 + @xmath76 & @xmath77 & @xmath78 & @xmath79 & @xmath93 & @xmath81 & @xmath82 & @xmath83 + 161.7 & 7.8 & 322.0 & 123.0 & 296.9 & 304.7 & 208.0 & 0.79 + & & & & + & & & & + & & & & + & +
our first goal is to look for parameter space in the nmssm having leptons as the main dark matter annihilation products .
this goal can be achieved if the lightest @xmath0-odd scalar is an intermediate annihilation product and subsequently decays to two taus or two muons from kinematic constraints .
the other goal is to have a large dark matter annihilation cross section to explain the size of the pamela excess .
the annihilation cross section can be enhanced through @xmath3-channel resonance effects . assuming the lightest neutralino is the dark matter candidate in the nmssm , the heavier @xmath0-odd scalar @xmath4 is the only particle which can play this role .
the @xmath0-even scalars are ruled out by the @xmath0 symmetry , because the initial state with two identical majorana fermions is @xmath0-odd .
the lighter @xmath0-odd scalar @xmath2 is far below the necessary mass region for the resonance effect . having established @xmath94 as the main dark matter annihilation channel ,
the main annihilation products are equivalent to the decay products of @xmath4 .
the main dark matter annihilation products are shown in fig .
[ fig : annihi ] for the two cases exemplified by table [ tab : case1para ] and table [ tab : case2para ] .
+ to calculate the dark matter annihilation cross section , we will concentrate on the main annihilation channel @xmath95 , which is true for both two cases .
the annihilation rate is & & ( a_2a_1h_1)v_dm + & = & ( 1- ) , + [ eq : anni_cross_section ] up to @xmath96 where the average dark matter speed is @xmath97 in the galactic halo . here
the mass of @xmath2 is neglected and the coupling @xmath98 has mass dimension one .
the co - annihilation effects can be neglected here , because other superpartner masses are much larger than the lsp mass . for the parameter points in table [ tab : case1para ] and table [ tab : case2para ]
, we have @xmath99 approximately 135 pb@xmath100c and 301 pb@xmath100c , respectively . also from table [ tab : case1para ] and table [ tab : case2para ] , the width @xmath83 of @xmath4 is below 1 gev , so the width part in the denominator of eq .
( [ eq : anni_cross_section ] ) can be neglected for @xmath101 gev considered here .
therefore , a large dark matter annihilation cross section can easily be obtained for the case of @xmath102 . on the contrary , if @xmath103 , the decay channel of @xmath104 is open and the decay width of @xmath4 is of order 10 gev .
then the dark matter annihilation cross section is limited by the width part in eq .
( [ eq : anni_cross_section ] ) , generically below 10 pb@xmath100c and not large enough to explain the pamela data .
thus we have _ two _ reasons to believe that the lightest neutralino mass is below the top quark mass : one is to have dominantly leptonic annihilation final states and the other one is to have a large annihilation cross section .
our annihilation rates are by far larger than the necessary one ( @xmath105 pb@xmath100c ) to satisfy the dark matter thermal relic density .
one possible explanation for this discrepancy is that the dark matter is nonthermal .
for example , other long - lived particles like the gravitino or moduli can decay into the lsp at a later time @xcite . as argued in @xcite , a long - lived modulus field with a lifetime shorter than one second can naturally appear in the anomaly - mediated susy breaking model .
the main dark matter relic abundance will be determined by the moduli annihilation cross section , which in principle can be smaller than the one for the neutralino and provide the observed dark matter relic density at the current time .
we now calculate the positron spectrum out of dark matter annihilation and compare it with the pamela results to determine the pamela - favored parameter space in the nmssm .
we first discuss the source term of primary positrons from dm annihilations , propagation of cosmic ray positrons and then the positron fluxes measured at pamela .
we also consider the charge - dependent solar modulation on the positron fraction spectrum .
for the two cases we are considering in this paper , the source term for primary positrons can be generally written as : q ( , e)=v ( ) ^2 . [ eq : positronsource ] here the overall factor @xmath106 is from the majorana property of the neutralino dm candidate ; @xmath107 is the thermally averaged annihilation cross section and to a good approximation can be replaced by the formula in eq .
( [ eq : anni_cross_section ] ) ; @xmath108 is the energy spectrum of positrons ; @xmath109 is the dark matter distribution inside the milky way halo . for the dark matter
distribution we use either the navarro , frenk and white ( nfw ) profile @xcite or the cored isothermal ( iso ) profile @xcite ; the use of other profiles may change discussions in this paper , especially for the gamma ray spectrum . the nfw and iso profiles are parametrized as : _
nfw(r)&= & _ ( ) ( ) ^2 , + _ iso(r)&= & _ ( ) , with @xmath110 kpc ( nfw ) , 5 kpc ( iso ) is the radius of the central core ; @xmath111 kpc is the galactocentric distance of the solar system ; @xmath112 gev @xmath113 is the solar neighborhood dm density .
the positron energy spectrum function for the @xmath5 case can be generally expressed as & = & de_a_1de_^+(e_a_1e_^+)(e_^+e_e^+ ) + & + & 2 de_h_1de_a_1de_^+(e_h_1e_a_1)(e_a_1e_^+)(e_^+e_e^+ ) .
[ eq : dndeformula ] here @xmath114 denotes the probability of a particle @xmath115 with energy @xmath116 decaying into a particle @xmath117 with energy @xmath118 ; the factor of 2 in the expression is because @xmath1 decays into 2 @xmath2 s .
we neglect the positrons and electrons from the @xmath62 quark decays although the @xmath1 mainly decays into two @xmath62 quarks in the muon - favored case . because positrons and electrons from @xmath62 decays are relatively soft , including them makes only a slight change in the spectrum below 10 gev , where the background is anyway dominant . from two @xmath35 s annihilating into @xmath2 and @xmath1 , and
neglecting the mass of @xmath2 , we have & = & br(a_2h_1a_1)(e_a_1-(m_- ) ) , + & = & br(a_2h_1a_1)(e_h_1-(m_+ ) ) . + since the emission of muons in the @xmath2 rest frame and the emission of @xmath2 s in the @xmath1 rest frame are isotropic , we have ( e_a_1e_^+)&&(e_a_1-e_^+ ) , + ( e_h_1e_a_1 ) & & , with @xmath119 is the heavy - side function and the masses of @xmath120 and @xmath2 are neglected in the approximation formula .
neglecting the muon polarization , the positron energy probability in muon decay has the following analytic form @xcite ( e_^+e_e^+ ) & = & + & & h(e_^+-e_e^+ ) , with @xmath121 . here the functions @xmath114 are normalized such that @xmath122 .
the @xmath7-favored case is more complicated than the @xmath5-case , because other than leptonic channels @xmath7 can also decay to various charged mesons , which decay eventually to electrons and positrons . in order to obtain the electron energy dependent probability from @xmath7
decays , we use pythia @xcite , which calls the program tauola @xcite , to simulate the inclusive electron / positron energy spectrum from both direct and indirect @xmath7 decays .
energy , respectively .
the blue line is from a fitted analytic function in the text .
lower panel : similar as the upper panel , but for electrons . on average , 1.16 positrons and 0.16 electrons are generated from one @xmath123 decay.,title="fig:",scaledwidth=45.0% ] + energy , respectively .
the blue line is from a fitted analytic function in the text .
lower panel : similar as the upper panel , but for electrons . on average , 1.16 positrons and 0.16 electrons are generated from one @xmath123 decay.,title="fig:",scaledwidth=45.0% ] as can be seen from fig .
[ fig : taudecay ] , once the energy of the ( unpolarized ) @xmath7 is much larger than its mass , the fractional electron and positron energy spectra are independent of the energy of the @xmath7 .
the large fluctuations for @xmath124 in the lower panel of fig .
[ fig : taudecay ] are due to limited monte carlo statistics .
the tiny fraction of @xmath125 out of @xmath123 for @xmath124 can be understood from the fact that only high multiplicity final states can contain an @xmath125 with a different charge from @xmath123 . for these @xmath126 and @xmath125 spectra from inclusive
@xmath123 decays , the following fitted functions provide good agreement : & & ( e_^+e_e^+)=e^-97.716x^5 + 223.389x^4 + & & e^-193.748x^3 + 82.595x^2 - 22.942x+2.783 ( 1-x ) , + & & ( e_^+e_e^-)=e^-15.575x^3 + 15.79x^2 + & & e^-18.083x+0.951(1-x ) , [ eq : taufragmentation ] with @xmath127 . for each @xmath123 , on average , there are @xmath128 positrons and @xmath129 electrons produced in the decay .
when @xmath2 decays into @xmath130 , there are on average 1.32 positrons out of a single @xmath2 decay .
therefore , we insert eq .
( [ eq : taufragmentation ] ) into the analogous formula in eq .
( [ eq : dndeformula ] ) for the @xmath7 case .
the propagation of positrons and electrons in the galactic medium is described by the following transport equation : -- [ b(e)n]=q ( , e ) , where @xmath131 denotes the positron number density per unit energy ; @xmath132 is the positron source term ; @xmath133 is the diffusion constant with @xmath134 gev ; @xmath135 is the positron energy synchrotron and inverse compton loss rate with @xmath136 s. the diffusive halo is modeled as a cylinder with radius @xmath110 kpc and the vertical direction @xmath137 inside @xmath138 .
the half thickness is not constrained and varies from 1 to 15 kpc .
we will consider three different parameter points for the cosmic ray propagation model in table [ tab : propagationparameter ] .
0.5 @ lcccc model & @xmath139 & @xmath140 [ kpc@xmath141/myr ] & @xmath142 [ kpc ] + m2 & 0.55 & 0.00595 & 1 + med & 0.70 & 0.0112 & 4 + m1 & 0.46 & 0.0765 & 15 + those sets of propagation parameters are compatible with the secondary / primary test for the secondary and primary antiprotons @xcite . assuming a time - independent state and considering the diffusion constant as space independent , we have -k_0()^n - = q ( , e ) .
[ eq : diffusion ] defining a pseudo - time and a relative pseudo - time between the source point and the observation point , respectively , as & & ( e)_e , + & & ( e , e_s)=(e)-(e_s ) , the characteristic diffusion length in the radial direction is _
d(e , e_s ) . in ref .
@xcite , an analytic solution for the positron flux on the earth has been obtained and has the following form : ^_e^+(e)&=&^_ed_e_sq(r _ , e_s ) + & & ( _ d(e , e_s ) ) . using the bessel expansion method , the halo integral @xmath143 , which is the volume integration of the green function of eq .
( [ eq : diffusion ] ) , has both radial and vertical expansions : ( _ d)=_i=1^_n=1^j_0(x_ir_/r_s)expr_i , n , with r_i , n=^r_s_0drrj_0(x_nr / r_s)^+l_-ldz()^2 .
here @xmath144 is the bessel function of the first kind and @xmath145 is the @xmath115-th root of the @xmath146 bessel function . in practice , one can use the numerically fitted functions in ref .
@xcite to speed up the numerical calculations .
for example , we use the following numerical function for the nfw dark matter profile and the m2 propagation model : ( _ d)&=&0.5 + 0.774 + & & , + with @xmath147 . pamela has measured the positron over electron fraction with energy below 10 gev , and obtained a spectrum significantly below the background fitted from other cosmic ray experiments .
one possible explanation of this discrepancy is due to the charge sign dependence of the solar modulation .
the magnetic field of the solar wind is dominated by the dipole term , and the projection of this dipole on the solar rotation axis can be either positive or negative , called @xmath148 and @xmath149 states , respectively . at each sunspot maximum , the dipole reverses its direction and leads to a periodic function for the dipole magnetic field with a roughly 12 year period . using two functions @xmath150 and
@xmath151 to model the solar modulation , we have the observed positron fraction on the earth as : f^_(e)= , with @xmath152 for the solar system in the @xmath148 state and @xmath153 for the @xmath149 state . here
@xmath154 denote the positron / electron fluxes .
one notices that only the ratio of @xmath155 is relevant for the positron fraction .
the ratio of the total electron flux in the @xmath148 cycle to total electron flux at a similar phase in the @xmath149 cycle is r(e)=. this ratio can be fitted in principle from the data shown in fig .
5 of ref .
however , since the lower energy bins of pamela ( where the contribution from dark matter is negligible ) are measured so precisely , one can use the data from pamela ( @xmath149 cycle ) and the known background @xcite ( without an assumption of solar modulation ) to determine this ratio .
the following fitted formula can achieve the goal : r(e)=max , which will be used in our following analysis about the comparison of model predictions and pamela data .
the general form of this expression is suggested by ref .
eventually one needs to understand or to calculate this ratio function or @xmath156 from first principles . due to the magnitude of the solar magnetic field ,
only positrons / electrons with energy below about 10 gev can be influenced by the solar wind .
@xmath157 is a monotonically increasing function and saturates at unity above @xmath158 gev . in terms of @xmath157 and @xmath159 , which is the positron fraction without solar modulation effects ,
one has : f^-_(e)&= & , + f^+_(e)&=&. before we compare the positrons from the neutralino annihilations in the nmssm to the pamela results , we first discuss the electron and positron backgrounds from standard astrophysicial processes .
background positrons are mostly secondaries originating from spallation processes of cosmic rays , mostly primary protons , off the interstellar gas , thus mainly occurring in the galactic disk .
the primary electrons are mainly produced by shock wave acceleration in supernovae . to simplify our comparison of model predictions to experimental data ( without the solar modulation effects ) , we use the following numerically fitted formulae for the background @xcite , which agrees with the full results calculated in @xcite : ^e^-_prim , bkg&= & , + ^e^-_sec , bkg&= & , + ^e^+_sec , bkg&= & , [ eq : electronbg ] in @xmath160 and with @xmath161 in gev . combining the background and signal positrons , we have a general formula for the positron fraction f^-_= = , assuming that the pamela data were taken when the sun is in the @xmath149 cycle . using the model point and the masses of @xmath35 , @xmath1 and @xmath2
reported in table [ tab : case1para ] , we have the positron excess for the @xmath5-favored model point in the nmssm shown in fig .
[ fig : muon ] . to generate the plot in fig .
[ fig : muon ] , we have used the m2 propagation model from table [ tab : propagationparameter ] , which provides a best fit to the pamela data .
the other two propagation models generate a flatter curve than the m2 model .
this is because as the thickness of the diffusive halo decreases , the positrons detected on the earth originate from a nearby region ( the characteristic propagation distance @xmath162 decreases ) , and hence low - energy positrons are less likely to reach the earth .
this leads to a steeper spectrum for the m2 propagation model .
since all three propagation models are supported by the n - body simulation , the combination of the nmssm and the pamela data ( assuming the nmssm interpretation was confirmed by , e.g. collider discoveries ) could help to determine a correct galaxy model . in fig .
[ fig : muon ] we note that the nmssm model is 3 sigma below the pamela data point in the highest energy bin .
however the overall agreement is quite good . to estimate the goodness of fit conservatively ,
we ignore the pamela data below 7.4 gev , which should have negligible contribution from the dark matter annihilation , and calculate the averaged @xmath163 for the 8 bins above 7.4 gev : = _
i=1 ^ 8()^20.9 . -favored model point in the nmssm .
the solid black line is the background without considering solar modulation .
the dotdashed black line is the background with the solar modulation effects .
the dashed blue line is the positron fraction from neutralino ( @xmath164 gev ) annihilations plus the modulated background .
the red points are data from pamela with one standard deviation errors .
the dark matter annihilation cross section is @xmath165 @xmath166 s@xmath167 .
the m2 propagation model is used here.,scaledwidth=48.0% ] if the neutralino annihilation in the nmssm is the explanation for the pamela data , the rising feature of the positron fraction spectrum should end at around 70 gev .
this is a dramatic prediction for future pamela results .
this predicted turnover follows uniquely from the requirement in the nmssm model that the dark matter lsp mass can not exceed the top quark mass . for completeness
, we also show the electron plus positron energy spectrum from dark matter annihilation in fig .
[ fig : muonelectron ] .
-favored model .
the solid black line is the background .
the dashed blue line is from the neutralino annihilation plus the background .
the red crossed points are the results from fermi large area telescope ( fermi lat ) with the gray band for systematic errors .
the dark matter annihilation cross section is @xmath168 cm@xmath141 s@xmath167 .
the m2 propagation model is used here.,scaledwidth=45.0% ] we also include the latest results from fermi lat in the red and crossed points @xcite in fig .
[ fig : muonelectron ] .
there is an additional uncertainty from the lat energy scale , which can shift the whole gray band by 5% ( up ) to 10% ( down ) and is not shown in this figure . in this figure the electron and positron fluxes from the background in eq .
( [ eq : electronbg ] ) plus signal have been normalized to agree with the first bin of fermi lat .
the positron fraction predictions from pamela are unchanged by this manipulation .
the agreement between the predicted electron+positron spectrum and the fermi lat data is about the same with and without adding the nmssm signal .
the generally poor agreement should thus be attributed to a defect in our understanding of the cosmic electron / positron background .
more generally , we conclude that a neutralino annihilation explanation for pamela is consistent with the fermi lat results , as long as the extra contribution to the electron+positron spectrum is within the fermi lat errors .
having discussed the @xmath5-favored point in the nmssm , we also report the results for the @xmath7-favored point .
it turns out that although @xmath7-favored points can be found easily in the parameter space of the nmssm , they provide a worse fit to the pamela data .
this is mainly because the positrons from @xmath123 decays are softer than the positrons from @xmath120 decays .
this fact can be seen from the lower panel of fig .
[ fig : tau ] , where a comparison between @xmath5 and @xmath7 cases is shown . -favored model point in the nmssm .
the dark matter ( @xmath170 gev ) annihilation cross section is @xmath171 cm@xmath141 s@xmath167 .
the m2 propagation model is used here .
lower panel : a comparison of positron energy spectra from @xmath123 decays and from @xmath120 decays.,title="fig:",scaledwidth=45.0% ] + -favored model point in the nmssm . the dark matter ( @xmath170 gev )
annihilation cross section is @xmath171 cm@xmath141 s@xmath167 .
the m2 propagation model is used here .
lower panel : a comparison of positron energy spectra from @xmath123 decays and from @xmath120 decays.,title="fig:",scaledwidth=43.0% ] only taking the pamela data above @xmath172 gev into account , we calculate the average @xmath163 for the fit of this @xmath7-favored model point to pamela data as @xmath173 . because of this large @xmath163 , we will concentrate on the @xmath5-favored model from here on .
since there is no excess of the antiproton fraction observed at pamela , this imposes constraints on the antiproton production cross section from dark matter annihilation . specifically to the muon - favored case in the nmssm , the dominant source of hadronic production is from @xmath174 and @xmath175 .
similarly to the calculations for the positron fraction spectrum , we first use an analytic function to fit the fragmentation function of @xmath176 to antiprotons .
the antiproton energy spectra for two different @xmath176 center of mass energies are extracted using pythia and shown in fig .
[ fig : bbarfrag ] . to antiproton fragmentation function .
the red and black points are for 20 gev and 300 gev @xmath176 center of mass energies .
the blue line is from a fitted analytic function described in the text .
the fragmentation function is extracted using pythia.,scaledwidth=46.0% ] we use the following fitted function later to replace the numerically simulated antiproton energy spectrum from @xmath1 decays : & & ( e_h_1b|bt_|p)= e^-307.05x^5 + 525.86x^4 + & & e^-331.51x^3 + 96.71x^2 - 28.56x+2.755 , where @xmath177 . as is customary in antiproton cosmic ray analyses , we substitute the proton energy by its kinetic energy @xmath178 in the following .
the antiproton spectrum from @xmath4 has a same formula replacing @xmath179 with @xmath180 .
the source term for antiprotons from dm annihilations is similar to eq .
( [ eq : positronsource ] ) but substituting @xmath181 with & & = de_h_1br(h_1b|b ) + & & ( e_h_1b|b t_|p)+de_a_2br(a_2b|b ) + & & ( e_a_2 b|b t_|p ) .
however , due to the fact that @xmath182 , the energy loss term for antiprotons can be neglected .
the steady diffusion equation for antiprotons is @xcite : -k_p(t)n & + & ( sign(z)v_cn ) + & + & 2h_annn = q ( , t ) .
[ eq : antiprotondiffusion ] here @xmath183 is the number density of antiprotons per unit energy and @xmath184 with @xmath185 and @xmath186 are the antiproton velocity and momentum .
the second term is related to the convective wind , which has a direction outward from the galactic plane and represents the movement of the medium responsible for the antiproton diffusion . 0.5 @
ccccc model & @xmath139 & @xmath140 [ kpc@xmath141/myr ] & @xmath142 [ kpc ] & @xmath187 [ km / s ] + m2 & 0.55 & 0.00595 & 1 & 13.5 + med & 0.70 & 0.0112 & 4 & 12 + max & 0.46 & 0.0765 & 15 & 5 + the velocity @xmath187 is assumed to be constant and has values shown in table [ tab : propagationparameter2 ] for different propagation models .
the third term represents annihilations of antiprotons and interstellar protons in the galactic plane with a thickness @xmath188 kpc .
the annihilation rate between antiproton and protons is @xmath189 with @xmath190 @xmath113 and @xmath191 @xmath113 .
the @xmath192 as a function of antiproton kinetic energy is given in @xcite and @xcite : @xmath193 mb for @xmath194 gev and @xmath195 mb for @xmath196 gev .
this annihilation process is dominant at low energy and leads to a decreased flux of antiprotons with low energy . in our analysis ,
other non - annihilation interactions between antiprotons and the interstellar medium are neglected .
these effects are not important for the antiproton flux with energy above a few gev .
similar to the positron case , the diffusion equation for the antiproton can be solved analytically and has the following concise form as its solution @xcite : ^_|p(t_|p)=()^2vr(t_|p ) . for the m2 propagation model and the nfw dark matter profile
, we numerically fit an analytical formula for @xmath197 r(t)/myr&=&10 ^ 1.352 + 0.0542_10(t)-0.265_10(t)^2 + & & 10^+0.0597_10(t)^3-,0.00575_10(t)^4 , + with @xmath198 in gev . to obtain the antiproton flux observed on the earth ,
we need to take the solar modulation effect into account .
the solar modulation effect is represented by a parameter @xmath199 , which is 500 mv for minimum solar activity when pamela was taking data .
the energy spectrum of antiprotons on the earth is ^_|p(t_|p)&=&^_|p(t_|p+|z| ) + & & , with @xmath200 the electric charge of the antiproton .
the background for primary protons can be extrapolated from other cosmic ray experiments . using the data from ams @xcite
, we arrive at the following fitted analytic function to describe the primary proton background : & & ^bg , ams_p(t)=e^-0.00097 ^ 5t+0.012 ^ 4 t + & & e^+0.014 ^ 3t-0.382 ^ 2t-0.828 + 6.88 , with @xmath198 in gev and the flux in m@xmath201s@xmath167sr@xmath167gev@xmath167 .
similarly , from caprice98 @xcite data , we have & & ^bg , caprice_p(t)=e^-0.0005 ^ 5t+0.005 ^ 4 t + & & e^+0.019 ^ 3t-0.433 ^ 2t-0.882 + 6.89 , the primary proton fluxes measured at ams and caprice can deviate from each other by a difference as large as 20% .
we include both results in our analysis to encode uncertainties of our primary proton flux background .
the secondary antiproton background can be found in the detailed analysis in @xcite and fitted in @xcite as : ^bg_|p(t)&=&10 ^ 0.028_10 ^ 4t-0.02_10 ^ 3 t + & & 10 ^ -1.0_10
^ 2t+0.07_10t-1.64 .
gev and its annihilation cross section is @xmath165 cm@xmath141 s@xmath167 . , scaledwidth=50.0% ] a comparison of the antiproton / proton flux ratio from the neutralino annihilation and the pamela data is shown in fig .
[ fig : bbar ] . noticing that the m2 propagation model provides a better fit to the pamela positron data , the model point in table [ tab : case1para ]
is marginally allowed by the pamela antiproton data . to quantify the discrepancy between the model prediction and
the pamela observed data , we calculate the @xmath163 by including all bins .
we also calculate the @xmath163 between the background and pamela data . using the background extrapolated from ams
, we have @xmath163 to be 1.6 for the dm prediction plus the background and 3.6 for the background only . while , using the background from caprice , we have @xmath163 to be 4.0 for the dm prediction plus the background and 2.2 for the background only .
we conclude that the model point in table [ tab : case1para ] is allowed by the pamela antiproton data , taking into account the large uncertainties of the background primary proton flux .
although the branching ratio @xmath202 is smaller than @xmath203 , the center of mass energy of @xmath176 directly out of @xmath4 is approximately twice of the center of mass energy out of @xmath1 .
therefore , the large antiproton fraction for the kinetic energy above 10 gev mainly comes from dark matter annihilating directly into @xmath176 . to suppress the antiproton flux more efficiently , one could change the model parameters to suppress the branching ratio of @xmath4 to @xmath176 .
one simple way to do this is to reduce @xmath73 .
however , by doing so , the branching ratio of @xmath204 is also increased , creating a tension with the upper limit from d0 described in section [ sec : tevatron ] .
if the @xmath5-favored model points in the nmssm are the correct explanation of the pamela positron excess , there will be lots of associated gamma rays generated . existing gamma ray data from hess @xcite and egret @xcite can in principle impose constraints on the dark matter annihilation cross section to electrons and positrons .
also , the recent and upcoming gamma - ray flux data from fermi lat have a smaller statistic uncertainty and can be used to test the pamela - favored nmssm model .
the differential gamma - ray flux from the dark matter annihilation has the following general formula : = ^_0 ^ 2(r)dl ( ) .
here @xmath205 with @xmath206 as the angle between the line of sight and the galactic plane .
one can separate the astrophysical uncertainties by introducing the quantity j()=^_0 ^ 2(r)dl ( ) . performing the solid angle integration ,
the differential gamma - ray flux is = |j ( ) , with @xmath207 and @xmath208 for the region around the galactic center .
for example , we have @xmath209 for @xmath210 and @xmath211 for @xmath212 sr using the nfw dark matter profile @xcite .
the gamma - rays from the @xmath5-favored model point of the nmssm have two sources : one is related to the muons in the final state and the other one is related to the bottom quarks in the annihilation final state .
altogether , we have & = & br(a_2a_1h_1 ) + & + & br(a_2a_1h_1 ) + & + & br(a_2b|b ) .
there are two processes to generate gamma - rays associated with the muon final state .
one is through final state radiation ( fsr ) and the other one is from the radiative muon decays into photons . for the final state radiation
, we have & & = ^_dx + & & ( ) ( ( ) -1 ) .
+ for the radiative muon decays : @xmath213 and @xmath214 , one has @xcite & & = + & & ^1_dx^min(1 , ) _
dyf(y ) . with @xmath215 and @xmath216 as the photon spectrum in the muon rest frame and given by ( y)&= & ( ( 3 - 2y+4y^2 - 2y^3 ) . + & & .
-+-+ . + & & .
+ ( 3 - 2y+4y^2 - 2y^3 ) ) , with @xmath217 .
the gamma - ray fragmentation function from bottom quarks in the dark matter annihilation final state is simulated using pythia and fitted using the following analytic function : & = & e^7.59 - 80.25x+412.5x^2 - 1297.7x^3 + & & e^+1969.7x^4 - 1137.8x^5 , with @xmath218 and @xmath161 as the center of mass energy of the @xmath176 system .
summing up all contributions to the gamma - rays , we compare the model predictions with the background for gamma - ray energy above 1 gev , which is fitted by a power - law in ref . @xcite as = 610 ^ -5()^-2.72 , in @xmath219 .
the fermi lat collaboration has already shown a preliminary result for the gamma rays from @xmath220 and @xmath221 @xcite .
gev and its annihilation cross section is @xmath168 cm@xmath141 s@xmath167 .
the solid green line is the dark matter prediction using the nfw profile plus the traditional background ( shown as the thin black line ) .
the dashed orange line is the dark matter prediction using the cored isothermal dark matter profile .
the dot - dashed green line is the dark matter prediction without including the contributions from @xmath62 quarks . , scaledwidth=50.0% ] here , @xmath222 and @xmath62 are the heliocentric galactic coordinates .
this region of angles corresponds to @xmath223 sr .
it is easy to calculate @xmath224 to be 13.4 for the nfw profile and 5.7 for the iso profile . in fig .
[ fig : fermilargeangle ] , we show the observed gamma ray fluxes from fermi lat together with the data from egret . as one can see , there is a disagreement between those two experiments for photon energy above 1 gev .
the predictions from dark matter annihilation using the model point in table [ tab : case1para ] tend to agree with the egret result if one uses the iso dark matter profile .
a larger discrepancy occurs for the dark matter prediction using the nfw profile .
this might be reconciled by astrophysical uncertainties like the smoothness of the dark matter distribution , which affects the necessary dark matter annihilation cross section by a factor of a few .
notice also in fig .
[ fig : fermilargeangle ] the dot - dashed green line showing the signal gamma - ray contribution from the muon final states only ; this indicates that the discrepancy is coming from the @xmath62 quarks in the final state .
we can also consider gamma rays coming from the galactic center . in fig .
[ fig : fermi ] , we compare the differential gamma - ray flux predicted from the neutralino annihilations in the nmssm to the flux from the background , after fitting the positron fraction spectrum of pamela .
gev and its annihilation cross section is @xmath168 cm@xmath141 s@xmath167 .
@xmath225 ( @xmath226 about the galactic center ) and @xmath227 ( nfw ) and @xmath228 ( iso ) are used here.,scaledwidth=50.0% ] we present the comparison of the model predictions and the background for a @xmath226 region about the galactic center . since the fermi lat will have an angular resolution of around @xmath229 ( @xmath230 sr ) , the error of their measurements in principle is small enough to test the dark matter annihilation scenario .
however , the model predictions for the gamma - ray flux in the galactic center are subject to large uncertainties from the dark matter profile , which can bring a factor of few hundred difference as seen from fig . [
fig : fermi ] . due to kinematical reasons ,
our dark matter neutralino mass is less than the top quark mass .
therefore , the gamma rays predicted in this model have an energy cutoff below the top quark mass .
hence , this model automatically evades the constraints from hess , which measured the gamma - ray flux with energy above 200 gev . for egret with @xmath231 around the galactic center and with the energy range of @xmath232 gev ,
the total background flux is @xmath233 cm@xmath201 s@xmath167 .
the dark matter contribution in this model is around @xmath235 cm@xmath201 s@xmath167 for the nfw profile and @xmath236 cm@xmath201 s@xmath167 for the iso profile .
this indicates a tension between the model prediction with the nfw profile and the egret data . again
this result is subject to astrophysical uncertainties ; for a solid angle around the galactic center , changing the dark matter profile can introduce an uncertainty of two orders of magnitude in the gamma - rays flux predictions .
there are many direct constraints on the nmssm parameter space from lep , tevatron , cleo , b - factories and the magnetic moment of the muon .
our analysis shows that the most stringent constraints are from exclusive @xmath237 decays at babar and light pseudoscalar searches at d0 .
compared to the mssm , the constraints from lep and other searches are less severe , due to non - standard decays of the mssm - like higgs boson and a new suppression factor @xmath238 beyond the mssm . the heavier bosons in the nmssm , typically above 250 gev , were unaccessible at lep .
therefore , we only consider the two lightest scalar particles @xmath2 and @xmath1 at lep .
there are three main production mechanisms for those two neutral higgs bosons .
one is through the higgsstrahlung process @xmath239 ; another one is the pair production process @xmath240 ; the third one is through the radiation off a massive fermion : @xmath241 . for the higgsstrahlung process @xmath239 and for an @xmath1 mass within the lep reach ,
the @xmath1 mainly decays to @xmath176 and @xmath242 for the @xmath5-favored and @xmath7-favored cases , respectively . for the @xmath7-favored model point
, there are many different final states like @xmath243 , @xmath244 and so on .
although the lep bounds on each of those channels are weak , the decay mode independent limits impose a bound on the mass of @xmath1 as @xmath245 gev @xcite , which is satisfied in the model points we have considered in this paper . for the @xmath5-favored case ,
the lower bound on the @xmath1 mass is around 114 gev , which is also satisfied for the model point reported in table [ tab : case1para ] .
the cross section of the pair production process @xmath240 is proportional to @xmath246 from eq .
( [ eq : a1h1z ] ) , and hence is tiny for @xmath247 and @xmath248 gev . for @xmath249 gev ,
the cross section is calculated to be @xmath250 fb for the model parameters in table [ tab : case1para ] .
considering the integrated luminosity of lep is below 1 fb@xmath167 , there are no constraints on the model parameters from this channel . finally , for the associated production with bottom quarks , the cross section is also suppressed due to a moderate @xmath73 and three - body final state phase space .
the cross section for @xmath241 at @xmath249 gev is calculated to be @xmath251 fb for the model parameters in table [ tab : case1para ] , which also indicates no constraints from lep .
the lightest @xmath0-even higgs boson @xmath1 has approximately the same couplings to fermions as in the standard model .
the ongoing searches at cdf and d0 do not yet constrain an @xmath1 with a mass around 115 gev at the tevatron @xcite . for the lightest @xmath0-odd higgs boson @xmath2
, the main production process at the tevatron is through associated production with @xmath176 and has a cross section proportional to @xmath252 , where @xmath199 has the same coupling to @xmath176 as the sm higgs boson .
although the production cross section can be large for the model parameters that we consider , the existing searches at d0 for the @xmath2 decays to two taus only constrain an @xmath2 with a mass above 90 gev @xcite . however , the recent searches for @xmath2 in the channel @xmath253 at d0 can impose a stringent bound on the muon - favored model parameter space in the nmssm .
the sm background for two pairs of collinear muons is very small ( below 0.02 events for @xmath254 fb@xmath167 integrated luminosity ) , and the null result imposes a constraint @xmath255 fb @xcite .
this bound is roughly independent of the @xmath2 mass .
the production cross section for a sm higgs at the tevatron is around 0.8 pb , so we need to have @xmath256 for @xmath257 . therefore , the muon - favored model point in table [ tab : case1para ] is allowed by this constraint , although it selects a specific region of the nmssm parameter space .
this experimental result points to small values of @xmath17 and @xmath19 to decrease the @xmath1 branching ratio into @xmath258 .
however , small values of @xmath17 and @xmath19 also increase the branching ratio of @xmath4 into @xmath176 .
since the dominant annihilation is mediated by the resonance effect with @xmath4 , this leads to a non - neglegible hadronic final state from the dark matter annihilation .
there is a new radiative contribution to the magnetic moment of the muon by exchanging @xmath2 in the loop diagram . using the one - loop result from ref .
@xcite , the new physics correction to @xmath259 is a_&=&()^2 + & & ^1_0 , and is negative .
the two - loop calculation will not change the sign of @xmath260 for a mass of @xmath2 below 1 gev @xcite .
the measured value of @xmath259 has a 3.4 @xmath261 deviation ( @xmath262 data only ) above the prediction of the standard model @xcite . requiring the new physics to be less than the experimental error
, we arrive at the following constraints on the model parameters : 2.5(bnl ( g-2)_/2 ) . here
@xmath77 is chosen to be 800 mev , while the constraints are less stringent as one increases @xmath77 .
notice that the @xmath260 may also receive significant contributions from other particles like smuons , that we chose to be heavy here to isolate the dark matter discussions .
another stringent bound on the nmssm parameter space with a light @xmath2 below 10 gev is from upsilon decays into a photon plus @xmath2 , which decays into a pair of taus or muons . for the mass range
@xmath263 gev , the strongest bound is from the recent cleo - iii limits @xcite on @xmath264 .
the radiative decay to @xmath265 is calculated as @xcite & = & ( ) ^2 + & & ( 1-)f , where @xmath266 incorporates qcd and relativistic corrections @xcite @xcite .
the data from cleo have the limit @xmath267 at 90% c.l . for a wide range of @xmath77 between @xmath268 gev to @xmath269 gev . using @xmath270 and @xmath271 from the model prediction
, this limit is translated into a bound on @xmath272 as 0.9 .
a similar result is obtained in @xcite . for the mass range @xmath273 gev ,
the strongest current bound is coming from the light scalar searches in the channel @xmath274 at babar . at 90% c.l .
, babar imposes an upper limit @xmath275
@xcite for @xmath77 below 1 gev . using @xmath276 and @xmath277 from the model prediction
, this limit is translated into a bound on @xmath272 as 0.4 .
[ eq : babar ] for the range of @xmath278 , the decay mode @xmath279 is open
. the branching ratio of this decay channel is given by @xcite & & b(k^+^+a_1)=(-^-1)^2 + & & ^2_a + & = & 310 ^ -6(-^-1)^2 ^ 2_a , which should be compared to the experimental values from the hypercp collaboration @xcite : @xmath280 .
therefore , the following constraint on the model parameter space is derived |-^-1|0.06 , so , other than when @xmath73 is very close to 1 or @xmath238 is extremely close to zero , the mass of @xmath2 is constrained to be above @xmath281 mev .
finally , there are also other constraints from b - physics like @xmath282 or @xmath283 . since the @xmath2 pseudoscalar does not mediate tree - level flavor - changing processes , one can use the minimal flavor violation assumption to suppress many kinds of flavor changing processes .
we have used the nmhdecay program to check those constraints and to justify the validity of our model points . in short , the most stringent constraint for @xmath284 gev is the radiative decays of upsilon into photons from babar .
the bound is @xmath285 .
one can see that the model point in table [ tab : case1para ] satisfies this bound .
the dark matter candidate neutralino in the nmssm from table [ tab : case1para ] is a combination of the bino , wino and higgsino .
therefore , it has a good chance to be detected in dark matter direct detection experiments , especially from the spin - dependent elastic scattering with nucleons .
here we just report the values calculated using the micromegas program @xcite .
the spin - independent dm - proton and dm - neutron cross sections are @xmath286 cm@xmath141 and @xmath287 cm@xmath141 , which are two orders of magnitude below the current bounds from xenon10 @xcite . for the spin - dependent one ,
the dm - proton and dm - neutron cross sections are @xmath288 cm@xmath141 and @xmath289 cm@xmath141 .
the later one is only one order of magnitude below the current bound @xcite and is in the accessible region of the upgraded experiments .
the nmssm explanation of the pamela positron excess can be tested by future experiments at colliders and new results from cosmic ray experiments . on the collider side , it is important to measure the masses of the neutralino and the heavier @xmath0-odd particle . if their masses satisfy the relation @xmath290 , the large dark matter annihilation cross section can be confirmed . another important quantity
to measure is the mass of the lighter @xmath0-odd particle , because the pamela positron excess prefers to have its mass below 1 gev .
it is also crucial to know the branching ratio of @xmath4 to @xmath176 , since the dominant antiproton contributions are from this channel . on the cosmic ray side
, we make a well - motivated unambiguous prediction that pamela will observe a turnover of the rising positron spectrum at around 70 gev .
the additional contributions to the electron+positron spectrum from dark matter annihilations are within current uncertainties but could be resolved in the future by fermi lat .
the associated gamma - ray flux from dark matter annihilation could also be resolved by fermi lat , but again due to astrophysical uncertainties one should be cautious when making a concrete prediction . in this paper
we have explored the possibility of using neutralino annihilations in the nmssm to explain the positron excess observed at pamela .
kinematics plays an essential role for having a viable model with a large fraction of leptons and a small fraction of hadrons in the annihilation final state .
the lighter @xmath0-odd particle @xmath2 has a mass below 1 gev and mainly decays into two muons .
the dark matter neutralino mass is less than the top quark mass to forbid the otherwise dominant @xmath10 final state . the @xmath3-channel resonance effect with the heavier @xmath0-odd particle @xmath4 increases the dark matter annihilation cross section to match the necessary one for the pamela positron fraction spectrum .
this also requires that the neutralino mass is less than the top quark mass , to avoid smearing the resonance effect from a large @xmath10 contribution to the @xmath4 width .
we have also shown that there is discrepancy between the nmssm predictions and the preliminary gamma ray fluxes from fermi lat . to alleviate this descrepancy
, one could evoke the existence of an astrophysical boost factor from a nearby clump of dark matter .
such boost factor would also improve the agreement between the nmssm prediction for the antiproton spectrum and the pamela results .
in particular , given the fact that positrons / electrons are only coming from nearby sources , a clump of dark matter can affect positron / electron , antiproton / proton and gamma ray fluxes differently , and would relatively increase the positron ratio spectrum more than the antiproton ratio spectrum and the gamma ray flux spectrum .
we have not utilized this astrophysical boost factor in our analysis .
an nmssm explanation of pamela makes three striking and uniquely correlated predictions : the rise in the pamela positron spectrum will turn over at around 70 gev , the dark matter particle mass is less than the top quark mass , and a light sub - gev pseudoscalar will be discovered at colliders .
* acknowledgments : * the authors are grateful to laura covi , gordon kane , maurizio pierini and peter skands for useful discussions .
fermilab is operated by fermi research alliance ,
llc under contract no . de - ac02 - 07ch11359 with the united states department of energy . _
note : after this paper was submitted in the arxiv , the babar collaboration presented a new constraint on the upsilon radiative decays to photon plus two muons by combining both @xmath291 and @xmath237 data @xcite .
this implies a more stringent bound on @xmath292 by a factor of two , see eq .
( [ eq : babar ] ) .
this can be accommodated in the nmssm by increasing the branching ratio of @xmath4 to @xmath176 .
this increases the tension for the antiproton ratio spectrum observed at pamela .
however , a modest astrophysical boost factor on the order of @xmath293 would loosen this tension .
_ o. adriani _ et al . _
[ pamela collaboration ] , nature * 458 * , 607 ( 2009 ) [ arxiv:0810.4995 ] .
o. adriani _ et al .
_ , phys .
lett . * 102 * , 051101 ( 2009 ) [ arxiv:0810.4994 ] .
m. cirelli and a. strumia , arxiv:0808.3867 ; v. barger , w. y. keung , d. marfatia and g. shaughnessy , phys . lett .
b * 672 * , 141 ( 2009 ) [ arxiv:0809.0162 ] ; i. cholis , l. goodenough , d. hooper , m. simet and n. weiner , arxiv:0809.1683 . m. cirelli , m. kadastik , m. raidal and a. strumia , nucl .
b * 813 * , 1 ( 2009 ) [ arxiv:0809.2409 ] .
n. arkani - hamed , d. p. finkbeiner , t. r. slatyer and n. weiner , phys .
d * 79 * , 015014 ( 2009 ) [ arxiv:0810.0713 ] ; m. pospelov and a. ritz , phys .
b * 671 * , 391 ( 2009 ) [ arxiv:0810.1502 [ hep - ph ] ] .
y. bai and z. han , phys .
d * 79 * , 095023 ( 2009 ) [ arxiv:0811.0387 ] ; p. j. fox and e. poppitz , [ arxiv:0811.0399 ] .
j. hisano , s. matsumoto and m. m. nojiri , phys .
lett . * 92 * , 031303 ( 2004 ) .
n. arkani - hamed and n. weiner , jhep * 0812 * , 104 ( 2008 ) [ arxiv:0810.0714 ]
. y. bai and z. han , arxiv:0902.0006 .
t. moroi and l. randall , nucl .
b * 570 * , 455 ( 2000 ) [ arxiv : hep - ph/9906527 ] ; b. s. acharya , p. kumar , k. bobkov , g. kane , j. shao and s. watson , jhep * 0806 * , 064 ( 2008 ) [ arxiv:0804.0863 ] ; m. pospelov and m. trott , jhep * 0904 * , 044 ( 2009 ) [ arxiv:0812.0432 ] .
p. grajek , g. kane , d. phalen , a. pierce and s. watson , arxiv:0812.4555 .
j. r. ellis , j. f. gunion , h. e. haber , l. roszkowski and f. zwirner , phys .
d * 39 * , 844 ( 1989 ) .
r. dermisek and j. f. gunion , phys . rev
. lett . * 95 * , 041801 ( 2005 ) [ arxiv : hep - ph/0502105 ] ; r. dermisek and j. f. gunion , phys .
d * 73 * , 111701 ( 2006 ) [ arxiv : hep - ph/0510322 ] ; r. dermisek and j. f. gunion , phys .
d * 79 * , 055014 ( 2009 ) [ arxiv:0811.3537 ] .
j. f. gunion , d. hooper and b. mcelrath , phys .
d * 73 * , 015011 ( 2006 ) [ arxiv : hep - ph/0509024 ] .
u. ellwanger , j. f. gunion and c. hugonie , jhep * 0502 * , 066 ( 2005 ) [ arxiv : hep - ph/0406215 ] .
c. panagiotakopoulos and k. tamvakis , phys .
b * 446 * , 224 ( 1999 ) [ arxiv : hep - ph/9809475 ] .
r. dermisek and j. f. gunion , phys .
d * 75 * , 075019 ( 2007 ) [ arxiv : hep - ph/0611142 ] .
y. nomura and j. thaler , arxiv:0810.5397 .
j. f. navarro , c. s. frenk and s. d. m. white , astrophys . j. * 490 * , 493 ( 1997 ) [ arxiv : astro - ph/9611107 ] . j. n. bahcall and r. m. soneira , astrophys .
j. suppl .
* 44 * , 73 ( 1980 ) .
t. sjostrand , s. mrenna and p. skands , jhep * 0605 * , 026 ( 2006 ) [ arxiv : hep - ph/0603175 ] .
s. jadach , z. was , r. decker and j. h. kuhn , comput .
commun . * 76 * , 361 ( 1993 ) .
t. delahaye , r. lineros , f. donato , n. fornengo and p. salati , phys .
d * 77 * , 063527 ( 2008 ) [ arxiv:0712.2312 ] .
f. donato , d. maurin , p. salati , a. barrau , g. boudoul and r. taillet , astrophys .
j. * 563 * , 172 ( 2001 ) [ arxiv : astro - ph/0103150 ] ; f. donato , n. fornengo , d. maurin and p. salati , phys .
d * 69 * , 063501 ( 2004 ) [ arxiv : astro - ph/0306207 ] .
m. cirelli , r. franceschini and a. strumia , nucl .
b * 800 * , 204 ( 2008 ) [ arxiv:0802.3378 ] .
j. clem , _ et .
_ astrophys .
j. * 464 * , 507 ( 1998 ) . i. v. moskalenko and a. w. strong , astrophys .
j. * 493 * , 694 ( 1998 ) [ arxiv : astro - ph/9710124 ] .
e. a. baltz and j. edsjo , phys .
d * 59 * , 023511 ( 1999 ) [ arxiv : astro - ph/9808243 ] .
fermi lat collaboration , arxiv:0905.0025 [ astro-ph.he ] .
l. c. tan and l. k. ng , j. phys .
g * 9 * , 227 ( 1983 ) .
r. j. protheroe , astrophys .
j. * 251 * , 387 ( 1981 ) .
j. alcaraz _ et al . _ [ ams collaboration ] , phys .
b * 490 * , 27 ( 2000 ) .
m. boezio _ et al .
_ , astropart .
* 19 * , 583 ( 2003 ) [ arxiv : astro - ph/0212253 ] .
t. bringmann and p. salati , phys .
d * 75 * , 083006 ( 2007 ) [ arxiv : astro - ph/0612514 ] . f. aharonian _ et al .
_ [ h.e.s.s .
collaboration ] , phys .
* 101 * , 261104 ( 2008 ) [ arxiv:0811.3894 ] .
s. d. hunter _
et al . _ , astrophys .
j. * 481 * ( 1997 ) 205 .
e. ponton and l. randall , jhep * 0904 * , 080 ( 2009 ) [ arxiv:0811.1029 ] .
r. essig , n. sehgal and l. e. strigari , arxiv:0902.4750 and references therein .
l. bergstrom , p. ullio and j. h. buckley , astropart .
* 9 * , 137 ( 1998 ) [ arxiv : astro - ph/9712318 ] .
n. giglietto , on behalf of the fermi lat collaboration , at la thuile 2009 .
m. krawczyk , arxiv : hep - ph/0103223 . g. w. bennett _ et al .
_ [ muon g-2 collaboration ] , phys .
d * 73 * , 072003 ( 2006 ) [ arxiv : hep - ex/0602035 ] ; m. passera , w. j. marciano and a. sirlin , aip conf .
* 1078 * , 378 ( 2009 ) [ arxiv:0809.4062 ] .
f. wilczek , phys .
lett . * 39 * , 1304 ( 1977 ) .
m. l. mangano and p. nason , mod .
a * 22 * , 1373 ( 2007 ) [ arxiv:0704.1719 ] .
i. g. aznaurian , s. g. grigorian and s. g. matinyan , jetp lett .
* 43 * , 646 ( 1986 ) .
f. domingo , u. ellwanger , e. fullana , c. hugonie and m. a. sanchis - lozano , jhep * 0901 * , 061 ( 2009 ) [ arxiv:0810.4736 [ hep - ph ] ] .
b. aubert [ the babar collaboration ] , arxiv:0902.2176 .
w. a. bardeen , s. h. tye and j. a. m. vermaseren , phys .
b * 76 * , 580 ( 1978 ) .
h. k. park _ et al . _
[ hypercp collaboration ] , phys .
* 88 * , 111801 ( 2002 ) [ arxiv : hep - ex/0110033 ] .
g. belanger , f. boudjema , a. pukhov and a. semenov , arxiv:0803.2360 .
j. angle _ et al . _
[ xenon collaboration ] , phys .
lett . * 100 * , 021303 ( 2008 ) [ arxiv:0706.0039 ] .
j. angle _ et al .
_ , phys .
lett . * 101 * , 091301 ( 2008 ) [ arxiv:0805.2939 ] .
b. aubert [ the babar collaboration ] , arxiv:0905.4539 [ hep - ex ] . | we examine whether the cosmic ray positron excess observed by pamela can be explained by neutralino annihilation in the next - to - minimal supersymmetric standard model ( nmssm ) .
the main dark matter annihilation products are the lightest @xmath0-even scalar @xmath1 plus the lightest @xmath0-odd scalar @xmath2 , with the @xmath2 decaying into two muons .
the energetic positrons needed to explain pamela are thus obtained in the nmssm simply from kinematics .
the required large annihilation cross section is obtained from an @xmath3-channel resonance with the heavier @xmath0-odd scalar @xmath4 .
various experiments constrain the pamela - favored nmssm parameter space , including collider searches for a light @xmath2 .
these constraints point to a unique corner of the nmssm parameter space , having a lightest neutralino mass around 160 gev and a very light pseudoscalar mass less than a gev .
a simple parameterized formula for the charge - dependent solar modulation effects reconciles the discrepancy between the pamela data and the estimated background at lower energies .
we also discuss the electron and gamma ray spectra from the fermi lat observations , and point out the discrepancy between the nmssm predictions and fermi lat preliminary results and possible resolution .
an nmssm explanation of pamela makes three striking and uniquely correlated predictions : the rise in the pamela positron spectrum will turn over at around 70 gev , the dark matter particle mass is less than the top quark mass , and a light sub - gev pseudoscalar will be discovered at colliders . |
the family of diluted magnetic semiconductors ( dms ) of the general form @xmath4 encompasses a wide variety of alloys which have been under extensive investigation during the past 15 years .
these alloys form a zincblende structure , where the @xmath5 element occupies one fcc lattice while @xmath6 and mn share the second fcc lattice .
one fundamental aspect of research has focused on the magnetic order of these systems , since they offer practical examples of strongly frustrated , randomly diluted three dimensional fcc heisenberg antiferromagnets ( afm ) with dominant nearest neighbour interaction@xcite . in this paper we present results of a monte carlo study of a diluted frustrated ising model on a fcc lattice given by the hamiltonian @xmath7 here , @xmath8 is the coupling constant , which we will set @xmath9 henceforward , and @xmath10 $ ] is the probability that a lattice site @xmath11 is occupied with an ising
spin @xmath12 .
we are interested in the static properties of this model for different dilution regimes . besides the pure ( @xmath13 ) and the slightly diluted case ( @xmath14 ) , that has already been studied by mc simulation@xcite and other methods@xcite , we concentrate our interest on the strong dilution regime , which has only been investigated in experimental heisenberg systems as mentioned above .
although in our work we perform a simulation of an ising system we find that some typical dms results can be reproduced with our simplified model .
neutron diffraction experiments of thin @xmath3 films for @xmath15 $ ] ) revealed a first order phase transition to an antiferromagnetically ordered state of `` type iii''@xcite . at
approximately @xmath16 a tricritical point is encountered , where the phase transition becomes continuous . the antiferromagnetic order in the regime
@xmath17 $ ] is truly long range , however , below @xmath18 a transition to a short range ordered phase seems to appear @xcite
. below @xmath19 most experimental results have led to the view , that one encounters a transition to a spin glass like phase at a fairly well defined temperature @xmath20 .
characteristic spin glass features are : ( i ) remanence effects in the frozen state@xcite , ( ii ) a pronounced cusp in the susceptibility around @xmath20@xcite with strong frequency dependence of the cusp temperature@xcite , ( iii ) absence of long range spin order as observed by magnetic neutron diffraction@xcite , ( iv ) dynamic scaling near @xmath20 of frequency - dependent response function@xcite and most importantly ( v ) a divergent nonlinear susceptibility around the cusp temperature@xcite .
on the other hand , the antiferromagnetic correlation length @xmath21 grows continuously with decreasing temperature until it saturates at the cusp temperature to an enormously large value as high as 70 at @xmath19 ; it is only below @xmath22 that short range afm order disappears . in the intermediate dilution range
@xmath23 the spin glass interpretation has been questioned , and it was suggested briefly@xcite ( on the basis of an `` activated '' scaling analysis ) that the equilibrium transition at @xmath20 was to the antiferromagnetically ordered state of a random field system .
this motivates a numerical study of a diluted antiferromagnet , in which we can observe the interplay of strong afm local order with spin glass order , and can measure the quantities now considered to be the signatures for spin - glass transitions .
because heisenberg model simulations demand more computer time , and because of the more convoluted controversies regarding the existence of a sharp phase transition for continuous spins , we adopt here an ising model ; in addition , we have retained only the first neighbour exchange interaction .
the paper is organized as follows : section [ sec : theory ] is concerned with the main theoretical arguments that guide our expectations for the results of our simulations in the distinct dilution regimes . in particular , we discuss the possible universality class of the proposed spin glass phase .
section [ sec : tech ] describes technical aspects of our simulations , in particular the equilibration criterion . in section [ sec : first - order - afm][sec : spin - glass - order ] of our report we shall present representative data for the distinct dilution regimes . in section [ sec : first - order - afm ] we concentrate on the pure case and on weak dilution ( @xmath14 ) , where we investigate how the order of the transition is being modified by disorder in form of stochastically removing spins from the lattice .
section [ sec : cont - afm ] is concerned with the regime @xmath24 , where a continuous phase transition with afm ordering is found ; critical exponents are determined by finite size scaling .
section [ sec : spin - glass - order ] investigates the intermediate and strong dilution range ( @xmath25 $ ] ) , where the question of the magnetic ordering is our main concern .
a summary of the results , a comparison to experimental systems and our final conclusions will be given in section [ sec : conclusion ] .
in this section , we collect the qualitative expectations for all the different concentration regimes expected in the phase diagram . as with the results in the subsequent sections , we begin by reviewing the pure case and proceed in the direction of greater dilution .
the global phase diagram is qualitatively similar to those conjectured for _ vector _ spins in ref . and ref . , except of course that distinctive collinear and noncollinear phases can not exist in the ising case
the pure ising antiferromagnet on a fcc lattice has been extensively studied both analytically and with simulations .
each spin has 12 nearest neighbours which in the ground state can only satisfy 8 bonds , 4 of them being always violated .
this effect of frustration , which follows from the triangles in the fcc lattice , leads to a large ground state degeneracy@xcite of the order @xmath26 , where @xmath27 denotes the linear extent of the system .
thus the entropy per spin is zero as @xmath28 . at small temperatures in this system ,
thermal fluctuations generate free energy terms which have the same effect as ferromagnetic second neighbor interactions : this favors the `` type i '' afm order , meaning that the system orders into one of the six periodic ground states with a @xmath29 type ordering wavevector@xcite .
this is an example of what villain called `` ordering due to disorder '' @xcite .
the discrete choice between the three @xmath29 type directions suggests a similarity in behavior to the 3-state potts model , which has a weakly first order phase transition in three dimensions .
indeed , the @xmath30 renormalization group predicts a first - order phase transition@xcite .
simulations @xcite and series expansions@xcite confirm that the phase transition is at finite temperature and is of first order .
the antiferromagnetic state may be handled quantitatively by constructing ( in the spirit of ref . )
a three - component staggered - magnetization order parameter @xmath31 with components @xmath32 for @xmath33 ; here the ordering wave vectors are @xmath34 and cyclic permutations ( we have taken the lattice constant to be unity . ) .
dilution in frustrated systems ( without any external field ) couples to the order parameter as a random field does in a ferromagnet@xcite .
take the case of rather weak dilution , which justifies assuming ( as a sort of variational state ) one of the six @xmath29 type ground states .
consider the effect of strengthening one _ bond _ lying within the @xmath35 plane : it will favor the four states with ( 100 ) and ( 010 ) ordering wavevectors and disfavor the two with ( 001 ) wavevectors , since the bond in question is violated in the ( 001 ) states .
the effect is much like a random field , except that it does not destroy the global up / down symmetry . in our case of _ site _ dilution ,
the random - field - like effects of removing an isolated site cancel each other ; however , removing a _ pair of sites _ has the same effect as would strengthening the bond between them@xcite . quite generally ,
when the random field is sufficiently strong , the first - order transition is converted to a continuous one@xcite . in the present context , since the effective random field grows with dilution , this argument predicts a tricritical point : the ordering transition is first - order for @xmath36 but becomes continuous for @xmath37@xcite .
for @xmath38 the transition from the paramagnet is expected to be a novel universality class@xcite . it would seem plausible if its dynamic scaling behavior were of the `` activated '' type , as in the random - field ising model .
no experiments have tested this , however ( the materials in this concentration range , roughly @xmath39 , can be grown only as thin epitaxial slabs , meaning that very little signal is available for susceptibility experiments . ) .
a claim was made that `` activated '' scaling could fit the data @xcite for lower values of @xmath40 , which we would identify as the spin - glass phase , but this was quickly corrected@xcite .
when the `` effective random fields '' are sufficiently strong , the antiferromagnetic order disappears and is replaced by spin - glass order at a critical value @xmath41@xcite . note that this threshold to propagate afm order is far above the @xmath42 , the geometrical percolation threshold for propagating connectivity of nearest - neighbor sites@xcite ( on the fcc lattice , @xmath43@xcite . ) .
any spin glass , by definition , requires random frustration .
this can be realized by dilution of a uniformly frustrated antiferromagnet , as in the present case , just as well as by a random mix of ferromagnetic and antiferromagnetic interactions@xcite .
indeed , the effective coupling between two spins may be ferromagnetic or antiferromagnetic depending on how intervening sites happen to be occupied .
of course , this spin glass state is expected to show residual short range correlations as opposed to the cases of the @xmath44 or gaussian random bond distributions , where symmetry implies that @xmath45=0 $ ] if @xmath46 .
it is intriguing that this `` spin glass '' state , which occurs below @xmath41 , from the viewpoint of the antiferromagnetic phase , is the same as the disordered domain state which is favored by the `` effective random fields '' mentioned above@xcite .
this state is different from the familiar random field disordered state ( and similar to the usual spin glass state ) because the global up / down symmetry is preserved ; consequently , for @xmath47 there is still a true phase transition in which this symmetry is broken@xcite . in a concentration range @xmath48
, the ground state is presumed to be a spin glass .
spin glass investigations tend still to be preoccupied with the issue of the existence of a transition as a function of dimension , external field , and spin type . indeed
, it is still unsettled whether the @xmath49 ising spin glass really has a transition at finite temperature , or whether it is at the lower critical dimensionality@xcite .
rather little has been done to test the universality of the critical exponents , as almost all simulations have used simple cubic ( sc ) lattices with the discrete @xmath44 distribution of random bonds .
monte carlo and series studies for the fcc lattice with @xmath44 bonds @xcite gave values of the spin - glass exponents @xmath50 , @xmath51 , and @xmath52 consistent with the sc @xmath44 model ; so did _ diluted _ @xmath44 model on a simple cubic lattice@xcite and ( modulo large error bars ) gaussian - distributed random bonds on the simple cubic lattice @xcite .
the above results are consistent with universality .
however , it has also been proposed that @xmath50 is more negative and the binder cumulant is larger for gaussian bond randomness than for @xmath44 randomness@xcite ; presumably the diluted fcc is more similar to the latter model , since its discrete randomness generically allows exact degeneracy of ground states .
we now consider the approach @xmath54 , where the spin glass long range order finally disappears . in this regime , the order is just barely propagating along tortuous , effectively one dimensional paths , and consequently we expect @xmath55 ( exponentially ) as @xmath54@xcite . note that @xmath56
. in frustrated models with discrete bond distributions , such as the present case , two portions of a connected cluster might be connected by ( say ) two chains of bonds , each canceling the other and allowing one portion to be flipped relative to the other portion at no cost in energy ; for propagation of order , it is as if no chain existed , i. e. , the effective concentration of bonds is lowered by the cancellations .
we use the single spin flip monte carlo metropolis algorithm in our simulations .
spins are updated sequentially and randomly .
periodic boundary conditions are imposed , limiting the possible lattice sizes to even numbers .
spins are represented on a cubic lattice with next nearest neighbour interactions to obtain a fcc lattice
. therefore , every lattice of linear size @xmath27 contains @xmath57 sites .
we simulated lattice sizes @xmath58 with @xmath59 realizations of the disorder and @xmath60 with @xmath61 .
we investigate the model in the concentration range @xmath62 $ ] .
a standard criterion by bhatt and young@xcite was applied to test the equilibration of the systems throughout the whole simulation , where we observe a continuous phase transition .
the procedure is to obtain two estimates of the spin glass phase indicator , the spin glass susceptibility @xmath63\ ] ] here and later , the brackets @xmath64 denote thermal averaging and @xmath65 $ ] configurational averaging .
we obtain @xmath66 by calculating the second moment of the spin glass order parameter defined in the two alternative ways , ( i ) as the overlap @xmath67 and ( ii ) as the autocorrelation ( self overlap ) @xmath68 here , @xmath69 and @xmath70 denote two sets of spins ( replicas ) with the same set of occupied sites and uncorrelated initialization and @xmath71 is the time initially used for equilibration . with these definitions ,
we can compute two estimates of @xmath66 as follows , i. e. , @xmath72,\ ] ] respectively the four spin correlation function , @xmath73,\ ] ] where @xmath74 , and @xmath75 .
the equilibration time @xmath71 was raised on a logarithmic scale and we only accepted a run , if both estimates of @xmath66 agreed after this time within certain limits , typically of the order of 5% of their joint mean value .
the longest runs performed were up to @xmath76 monte carlo steps per spin ( mcs ) .
most of the simulations were performed on hp workstations at the institut fr theoretische physik , gttingen and on the intel paragon parallel computer at the hchstleistungsrechenzentrum jlich .
the program was parallelised by using _
pvm 3.2 _ software @xcite in order to simulate many systems simultaneously .
in this section we investigate the pure afm and the slightly diluted regime , i. e. , @xmath77 $ ] .
we wish to determine the order of the phase transition and how the order of the transition is changed by introducing disorder into the system in form of slight stochastic dilution .
the magnetic order in this regime is clearly antiferromagnet , consistent with earlier simulations , and shall be closer examined in section [ sec : cont - afm ] .
the pure antiferromagnet on a fcc lattice is known to undergo a temperature driven first order phase transition@xcite , as mentioned in the previous section .
early monte carlo simulations by grest and gabl@xcite as well as giebultowicz@xcite reported a change from a first order to a continuous phase transition upon dilution .
grest and gabl located the tricritical point at a critical concentration @xmath78 using ising spins , whereas giebultowicz found a slightly lower @xmath79 with a heisenberg spin simulation .
however , in both of those simulations , no averaging over the disorder was performed , so that we re investigated this regime .
an important quantity for a first order phase transition is the internal energy density @xmath80=\left[\left\langle\frac{1}{2n}\sum_{<i , j>}\epsilon_i \epsilon_j s_i s_j\right\rangle\right].\ ] ] at the critical temperature , this quantity indicates a first order phase transition by a discontinuity ( latent heat ) , which can be seen in fig .
[ energy - heat ] , where @xmath81 $ ] is plotted versus temperature @xmath82 ; with increasing lattice size a pronounced discontinuity at the transition temperature can be observed , revealing clearly a first order transition . in fig .
[ energy - p=0.9 ] we present our data of the internal energy density for @xmath83 . for this concentration
the transition appears to be continuous .
apparently , a tricritical point is hard to locate with these monte carlo data , since it is not clear , whether in the limit of smaller temperature steps and larger systems the transition will turn out to be continuous or not . in order to determine the tricritical concentration
more efficiently we used a method first introduced by lee & kosterlitz @xcite , which is based on a histogram algorithm by ferrenberg & swendsen@xcite .
this method@xcite uses one long monte carlo run to estimate the free energy at several temperatures close to @xmath20 . during the runs a histogram of the internal energy density @xmath84 is accumulated .
this histogram serves as an estimate of the equilibrium energy distribution ( after correct normalization ) @xmath85 here , @xmath86 is the partition function at the inverse temperature @xmath87 and @xmath88 is the entropy .
notice , that @xmath89 ( [ ener - dist ] ) is proportional to @xmath90 , where @xmath91 denotes the free energy .
the distribution @xmath89 can now be used to generate the distribution ( and consequently @xmath92 ) at a different inverse temperature @xmath93 in the vicinity of @xmath94 .
we are interested in the situation in which @xmath92 has two minima @xmath95 and @xmath96 ( i.e. ( [ ener - dist ] ) has maxima at these energies ) .
the temperature at which @xmath97 is taken to be the effective critical temperature for the given size @xmath27
. then we evaluate the `` gap '' @xmath98 between the free energy values at those two energies and at the maximum @xmath99 in between them . from a scaling analysis of @xmath100
, one can identify whether a phase transition is first order@xcite . a state with @xmath101 $ ]
consists of a domain of ordered and a domain of disordered phase coexisting , separated by a @xmath102-dimensional interface surrounding the droplet of minority phase .
therefore , one can expand the free energy @xmath103 where the bulk free energy density @xmath104 is minimal and constant for @xmath101 $ ] and the surface term @xmath105 is maximal at @xmath106 .
expansion ( [ free - ener - exp ] ) is valid for any first order phase transition , as long as the correlation length @xmath107 . at the critical temperature @xmath108
remains finite , so that for sufficiently large @xmath27 the appearance of a free energy gap @xmath100 indicates a first order transition ; for small @xmath109 the free energy is dominated by the bulk term .
as the system approaches a tricritical point , @xmath108 grows and the double minima structure can only be seen for large @xmath27 . at a tricritical point and beyond it , the phase transition is continuous and hence there is no double minimum structure for any @xmath27 . to obtain good statistics
, we took histogram data every @xmath110 mcs for @xmath111 mcs , averaging over 16 realizations of the disorder .
[ hist-2temp ] shows two histograms of the internal energy density for @xmath13 after normalization for two different temperatures ( full and dashed line ) .
the dots were produced by transforming the @xmath112 data set ( dashed line ) to the lower temperature @xmath113 .
the distribution depends sensitively on the chosen temperature ; the monte carlo data at @xmath113 and the transformed data from @xmath112 show good agreement , indicating , that the equilibrium distribution has been well estimated .
the simulation temperature in fig .
[ hist-2temp ] was chosen very close to the critical temperature . therefore , as noted above , the distribution clearly has two maxima at @xmath95 and @xmath96 , which are equivalently minima of @xmath92 .
if we define the effective critical temperature @xmath114 by the equality of the two maxima @xmath115 and @xmath116 and extrapolate @xmath117 , we obtain @xmath118 .
this is slightly lower than the values from ref .
( @xmath119 ) and ref .
( @xmath120 ) .
this discrepancy may be due to using different definitions of `` finite
size @xmath20 '' in the respective references .
while for @xmath13 the energy gap appears already at @xmath121 , for stronger dilution it can only be seen at larger system sizes , indicating the growing correlation length as one approaches the tricritical point .
if a gap appeared at one system size , it continued to be present and in fact grew for larger sizes .
the scaling behaviour according to eq .
( [ free - ener - exp ] ) , i. e. , @xmath122 , is fulfilled within the error margins . in the diluted case we found considerable fluctuations of the energy gap size depending on the realization . at @xmath123 for our largest systems ( @xmath124 ) certain realizations were found , that showed none of the typical two peak structure , while others still showed a small gap . in order to determine the tricritical dilution @xmath125 , we evaluated the percentage of realizations with gap ( @xmath126 ) with decreasing concentration of spins . for our largest systems ( @xmath124 ) , at @xmath127 @xmath128 of the realizations still showed a double peak , at @xmath83 only 40% , at @xmath129 10% and for @xmath130 no double peaks were found at all , i.
e. , all systems showed gaussian peaks at @xmath20 . from extrapolating @xmath131 determined by the condition @xmath132 for different system sizes ( see dots in fig .
[ phase - diag ] ) , we estimated @xmath133 in accordance with ref . .
to determine the triciritcal concentration more exactly would require a precise scaling analysis based on more extensive data , which was beyond the scope of this work .
in this section we concentrate on simulations performed for @xmath24 .
our aim is to check , whether at @xmath20 the system orders into an antiferromagnetic state and , since we are below the tricritical concentration we measure critical exponents via finite size scaling close to the continuous phase transition , that we encounter . for the antiferromagnetic phase , the staggered magnetization `` vector '' @xmath134 ( see eq .
[ staggmag ] ) is the appropriate order parameter .
thus , we calculate the second moment @xmath135.\ ] ] for @xmath136 this is proportional to the staggered susceptibility @xmath137 , which we analysed by using the finite size scaling form @xmath138 to extract the critical exponents @xmath50 and @xmath51 and the critical temperature @xmath20 .
since we are mainly interested in the magnetic order of the different dilution regimes , we also calculated correlation functions . to save computer time , we fourier transform the lattices to @xmath139space and calculate the fourier transformed correlation function , i. e. , @xmath140= \frac{1}{n}\sum_{ij } \exp ( -{\rm i}\ : \mbox{\small\boldmath $ r$}_{\!ij } \cdot\mbox{\footnotesize\boldmath $ k$})\ : [ \langle s_i s_j\rangle ] .\ ] ] we applied the fast fourier algorithm to the ( most relevant ) @xmath60 systems only and calculated the correlation function along the three @xmath141 directions . for an antiferromagnet , the three @xmath142 modes of @xmath143 should yield the static staggered susceptibility , which we used as a consistency check .
the correlation length @xmath108 can be extracted from the knowledge of the scaling form of @xmath143 , i. e. , @xmath144 the scaling factor @xmath145 has the following asymptotics : for @xmath146 and @xmath147 , @xmath148 and and for @xmath136 and @xmath149 ( in the afm case @xmath150 with the limit @xmath151 ) @xmath152 in eq .
( [ cor - scal - susc ] ) we have used the scaling form of the correlation length @xmath153 and the scaling law @xmath154 . with this knowledge of the asymptotic form of @xmath143 we chose the ansatz @xmath155 that we used to fit our data of the correlation function in order to obtain an estimate for @xmath108 and @xmath50 .
the constant @xmath156 has been introduced as the amplitude of the correlation function .
finally we shall analyse the heat capacity @xmath157.\ ] ] this quantity indicates a continuous phase transition to an ordered state by a weak divergence at the critical temperature .
we use the finite size scaling form @xmath158 to extract critical exponents @xmath159 and @xmath51 as well as @xmath20 . our data of the staggered magnetization show an increasing @xmath160 ( eq . [ stag - mag ] ) as @xmath82 approaches the critical temperature , becoming more pronounced for larger lattice sizes .
this contribution to the afm order parameter can as well be seen in the divergent behaviour of the staggered susceptibility ( fig .
[ stag - sus ] ) .
the scaling analysis@xcite of the staggered susceptibility yields @xmath161 , @xmath162 and @xmath163 .
the data scale well over a wide temperature range with the exception of the @xmath164 data and the errors in the exponents are within acceptable limits .
the antiferromagnetic correlation length @xmath21 is found to grow continuously with decreasing temperature and reaches half the lattice size at @xmath165 , before the critical temperature @xmath166 .
since finite size effects become appreciable at distances close to half the lattice size , we only included data above @xmath167 for a fit of the scaling form @xmath153 .
we extracted @xmath168 by regression , which is in good agreement with our previous result . from fitting the correlation function to eq .
( [ cor - fun - fit ] ) , we were also able to extract a second estimate of @xmath50 ; here , we found @xmath169 with a slow drift to @xmath170 for @xmath171 .
we do not perform a complete finite size scaling analysis of the correlation function in this work , because it would require a substantial amount of additional data for larger lattice sizes .
therefore , our estimates for the exponents obtained from @xmath172 are less reliable than those obtained from finite size scaling .
nevertheless , this analysis serves as an additional consistency check and the critical exponents lie well within the error margins of those exponents obtained via finite size scaling of the susceptibility .
the heat capacity shows a weak divergence at the critical temperature at @xmath24 .
[ heatp=8-scale ] displays the best result of the scaling procedure according to eq .
( [ heat - scale ] ) with @xmath173 , @xmath174 and @xmath175 .
these exponent values satisfy the hyperscaling relation @xmath176 with @xmath177 in 3 dimensions . in summary ,
the dilution regime @xmath24 exhibits a continuous phase transition to an antiferromagnetically ordered state .
we obtained critical exponents using finite size scaling ; however , since @xmath24 is very close to the tricritical point , it is quite possible that these are only effective exponents from the crossover between the mean field exponents of the tricritical point ( @xmath178 , @xmath179 and @xmath180 ) to whatever universality class is appropriate for the continuous ordering transition . with the data at hand this question has to remain open .
upon further dilution we encounter a dramatic change in the magnetic order of the system at the phase transition .
we investigate the question , whether below @xmath24 the system really orders into a spin glass or if antiferromagnetic ordering can still be found .
first , we introduce the spin glass order parameter and the spin glass susceptibility .
then we discuss our data at @xmath19 , and proceed with results from simulations of stronger diluted systems .
besides the quantities that have been introduced in the previous section , we also measured the standard indicator of a spin glass transition , the spin glass susceptibility @xmath66 ( see eq .
( [ chi - sg ] ) ) . for an infinite system
a spin glass transition is signalled by a divergence of @xmath66 as @xmath181 , with @xmath182 .
for our scaling analysis , @xmath66 is computed as the second moment of the overlap , defined by eq .
( [ overlap ] ) and eq .
( [ chi - sg - overlap ] ) .
we analysed our estimate of the spin glass susceptibility by using it s finite size scaling form @xmath183 another important quantity , that is well known in the analysis of spin glass simulation data@xcite , is the binder cumulant@xcite of the spin glass order parameter @xmath184}{\left [ \langle q^2\rangle\right]^2}\right).\ ] ] it has the pleasant finite size scaling form @xmath185 with no power of @xmath27 multiplying @xmath186 , which makes it very valuable for precise scaling analysis .
the binder cumulant ( [ binder - cum ] ) is defined so that @xmath187 , and above @xmath20 , @xmath188 for @xmath189 . in particular , the intersection of all @xmath190 curves at some point provides an accurate estimate of @xmath20 . to investigate a change of magnetic order further
, we analysed the correlation function again , this time also the spin glass correlation function @xmath191= \frac{1}{n}\sum_{i , j}\exp(-{\rm i}\ , \mbox{\small\boldmath $ r$}_{\!ij}\cdot \mbox{\footnotesize\boldmath $ k$})\ : [ \langle q_i q_j\rangle ] .\ ] ] here , we defined @xmath192 .
the same fitting procedure of our data was employed as in the previous section in order to compute the respective correlation length , @xmath51 and @xmath50 . the ( antiferromagnetic ) staggered susceptibility as shown in fig .
[ stag - susc=7 ] , is drastically reduced in comparison to fig .
[ stag - sus ] , and shows only a small tendency to increase as @xmath82 decreases .
furthermore , scaling according to eq .
( [ chi - dagg - scale ] ) could not be achieved for reasonable parameters . on the other hand
our data reveal a divergence of @xmath66 at @xmath19 of the same order of magnitude as @xmath193 for @xmath24 in fig .
[ stag - sus ] .
this divergence becomes particularly strong for larger lattice sizes in the vicinity of @xmath194 .
if @xmath66 is the critical quantity for this system then it should also scale according to eq .
[ scale - chi - sg ] . a finite size scaling analysis of the spin glass susceptibility
is given in fig .
[ chi - sgp=7-scale ] .
all data for @xmath66 with the exception of @xmath164 scale well with @xmath195 , @xmath196 and @xmath197 .
the fact that the spin glass susceptibility satisfies the above scaling form ( [ scale - chi - sg ] ) strongly suggests that the magnetic order of this model has changed between @xmath24 and @xmath19 from antiferromagnet to spin glass . in fig .
[ binp=7 ] we present our data of @xmath198 at @xmath19 .
the data show the typical behaviour as it has been observed in short range spin glasses , i. e. , the data merge at approximately @xmath194 , indicating the phase transition .
there is even a slight tendency of fanning out of the data below this intersection point , which is a strong evidence for the occurrence of a phase transition .
such a fanning out is usually observed in uniform systems . in our system
this may be due to the proximity to the tricritical point or to residual short range antiferromagnetic order ( see below ) .
the best fit was achieved with @xmath199 and @xmath200 , which agrees well with our estimates from the scaled spin glass susceptibility .
the stronger scattering of this quantity , especially of the @xmath201 $ ] data , may be due to insufficient disorder averaging ( 40 systems for @xmath60 ) and could probably be decreased with additional computational power . in fig .
[ sg - cor - len ] the temperature dependence of @xmath202 and @xmath21 are presented .
while the spin glass correlation length is very small for large @xmath82 but increases drastically as @xmath203 , the antiferromagnetic correlation length starts at a higher level but increases slower than @xmath202 ; it ceases to increase at about @xmath204 and saturates for lower temperatures .
all these results of our simulations at @xmath19 are consistent with the interpretation that in this dilution regime we witness indeed a spin glass transition , but also encounter antiferromagnetic order of long but _
finite _ range being embedded into long range spin glass order .
apparently , a change of magnetic order has taken place within the interval @xmath205 , where @xmath41 denotes the critical dilution , where this change happens .
additional simulations were performed for concentrations @xmath206 . with stronger dilution
we find that the spin glass phase already encountered for @xmath19 persists .
both , spin glass susceptibility and binder cumulant of the order parameter scale well in this regime with slight dilution dependent critical exponents .
also , in this regime the data of the binder cumulant stay together below @xmath20 for all sizes , as has been observed in other short range ising spin glass simulations .
unfortunately , it is increasingly difficult in this regime to equilibrate the systems . as the concentration is lowered , the critical temperature decreases rapidly ( as expected from our argument earlier , that @xmath207 with @xmath208 ) , but the characteristic microscopic energy barriers remain of order unity .
metastability becomes an increasingly important problem and is prohibitive in large systems . in our simulations , for @xmath209 we were unable to equilibrate the @xmath60 systems sufficiently close to @xmath20 within reasonable computer time .
from analysing the correlation function in the strong dilution regime we find two results : first , the spin glass correlation length shows a similar divergence close to the critical temperature for @xmath210 as we have seen in the previous subsection , confirming again the development of long range spin glass order .
second , we still find short range antiferromagnetic order , which decreases with lower concentration : for @xmath211 @xmath21 rises slowly when @xmath82 is lowered and at @xmath20 we have @xmath212 ; for @xmath22 the correlation length remains constant at @xmath213 for all temperatures which we can simulate ( @xmath214 , t_c(p=0.4)\approx 0.47 $ ] ) . furthermore , both critical exponents @xmath51 and @xmath50 apparently decrease with increasing dilution ( see the fitted values in table i ) .
the generic theoretical expectation is that the exponents should be universal all along the spin glass transition line , but the issue of universality is not completely settled in diluted systems@xcite . in our case , we note that the drift in the exponents is pronounced for @xmath215 but minimal for @xmath216 .
thus we suggest that this dependence is an artifact of the antiferromagnetic correlation length , which is large and @xmath40-dependent for @xmath215 , but is quite small for @xmath217 . however , we have insufficient information to check this proposition , and shall come back to this issue in the conclusion .
we conclude , that for the whole range @xmath25 $ ] our simulations testify the existence of a spin glass phase transition for the short range antiferromagnetic ising model on a fcc lattice .
simultaneously , the model also exhibits residual antiferromagnetic order , that is relatively long ranged at @xmath19 and saturates close to the critical temperature ; it decreases with further dilution , showing no temperature dependence for @xmath22 .
in this paper we have presented a monte carlo simulation of the diluted short range antiferromagnetic ising model on a fcc lattice . the main purpose was to investigate the thermodynamic equilibrium properties of this model .
in the undiluted case ( @xmath13 ) we have found a first order phase transition to an antiferromagnetically ordered state , consistent with earlier simulations . upon slight dilution ,
the antiferromagnetic order persists , the first order transition becoming weaker and changing to a continuous transition at the tricritical concentration @xmath218 . at @xmath24
we still find antiferromagnetic order from analysis of the correlation function and scaling of the staggered susceptibility .
together with scaling of the heat capacity , we find mutually consistent critical exponents .
the question of universality remains unclear ; in view of the proximity to @xmath218 it seems likely that we observe tricritical exponents at this point .
below @xmath24 the divergence of the spin glass correlation length as well as scaling of the spin glass susceptibility and the binder cumulant of the order parameter signal spin glass order .
this means , that there must be a multicritical point at some concentration @xmath205 , where the change of magnetic order takes place .
simultaneously , the quasi temperature independent staggered magnetization and the saturation of the antiferromagnetic correlation length suggest the breakdown of antiferromagnetic long range order .
the fact , that the antiferromagnetic correlation length saturates while spin glass correlations still grow contradicts the view of a dynamically inhibited transition to an antiferromagnet .
rather , our data suggest that the coexistence of antiferromagnetic short range order together with long range established spin glass order seems to be a special phenomena of this diluted model .
let us now turn to the question of universality in the spin glass phase .
in this phase ( see sec . [
sec : spin - glass - order ] ) our scaling fits yielded critical exponents slightly dependent on dilution ( compare table [ exponents ] ) ; we speculated , that this could be an artifact of the changing antiferromagnetic correlation length .
even in unfrustrated models , there is still controversy over the universality of exponents under dilution@xcite . although our simulations are less extensive than the mc
simulations of the short range ising spin glass by bhatt and young and by ogielski , it is interesting to compare our exponents with their values for the @xmath44model in @xmath49 . for this model ,
they find@xcite @xmath219 and @xmath220 .
these values are just at or slightly out of the error margins of the present simulations , so that it is not entirely clear , whether the two models lie in the same universality class .
note also , that our value of @xmath52 is smaller than for the `` classic '' spin glass simulations , roughly @xmath221 for @xmath25 $ ] .
additional simulations of larger systems would be desirable to confirm our result .
it should be noted that even the best current simulations on hypercubic lattices have not put to rest the basic question whether the lower critical dimension for ising spin glasses is below 3 , or equal to 3@xcite . at the more modest scale of our simulation , it can not be decisively answered whether the diluted fcc ising antiferromagnet has a genuine spin glass transition , and if so whether it has precisely the same exponents as the standard example of @xmath44 spin glasses on hypercubic lattices using array processors .
however , our data is consistent with both of these propositions .
although in this work we have performed an ising model simulation we find striking similarities with experimental observations in dms , which have heisenberg like magnetic moments@xcite . not only are the respective phase diagrams ( compare fig .
[ phase - diag ] and ref . ) qualitatively similar and the various critical concentrations are numerically close , but also the concentration dependence of @xmath21 in the spin glass phase agrees well with experiment . to be more specific
, a comparison of our simulations with intensive experimental studies by giebultowicz et al.@xcite on @xmath222 shows the following agreements : ( i ) the afm transition is first order for @xmath77 $ ] , ( ii ) the transition is continuous ( to an afm long range state ) for @xmath17 $ ] , ( iii ) the afm order is of large but finite range for @xmath223 in the spin glass phase and ( iv ) the afm correlation length decreases with increasing dilution @xmath224 .
these results together with the measured divergence of the nonlinear susceptibility in @xmath225 ( see ref . ) below @xmath226 do support the view of a spin glass phase in experimental heisenberg systems . c. w. and a. z. would like to thank reiner kree for helpful discussions and j. holm for his assistance .
c. w. and a. z. gratefully acknowledge support by sfb 345 . c. l. h. is supported by nsf grant dmr9214943 .
cdddd + order & concentration & susceptibility & binder cumulant & correlation function + & & @xmath20=1.07 ( 0.05 ) & & + afm & 0.8 & @xmath51=0.51 ( 0.10 ) & & @xmath51=0.55 + & & @xmath50=0.05 ( 0.15 ) & & @xmath2270.04 ( 0.05 ) + & & @xmath20=0.83 ( 0.05 ) & @xmath20=0.83 ( 0.05 ) & + sg & 0.7 & @xmath51=1.00 ( 0.20 ) & @xmath51=1.05 ( 0.20 ) & @xmath51=0.96 + & & @xmath50=0.10 ( 0.20 ) & & @xmath50=0.15 ( 0.20 ) + & & @xmath20=0.75 ( 0.05 ) & @xmath20=0.76 ( 0.05 ) & + sg & 0.6 & @xmath51=0.80 ( 0.20 ) & @xmath51=0.90 ( 0.20 ) & @xmath51=0.98 + & & @xmath50=0.00 ( 0.20 ) & & @xmath50=0.07 ( 0.15 ) + & & @xmath20=0.55 ( 0.05 ) & @xmath20=0.53 ( 0.05 ) & + sg & 0.5 & @xmath2280.73 ( 0.20 ) & @xmath2280.73 ( 0.20 ) & + & & @xmath2270.30 ( 0.25 ) & & + & & @xmath20=0.47 ( 0.10 ) & @xmath20=0.44 ( 0.10 ) & + sg & 0.4 & @xmath51=0.80 ( 0.20 ) & @xmath51=0.80 ( 0.20 ) & + & & @xmath2270.35 ( 0.20 ) & & + & & @xmath2290.27 ( 0.10 ) & @xmath2290.27 ( 0.10 ) & + sg & 0.3 & @xmath2280.70 ( 0.20 ) & @xmath2280.70 ( 0.25 ) & + & & @xmath2270.40 ( 0.20 ) & & + a renormalization - group approach could be formulated with a landau ginzburg effective free energy as a function of the order - parameter field @xmath134 . the fact that , in a particular ordered configuration , only one of the ordering wavevectors can be present , translates to a `` cubic '' anisotropy in order
parameter space favoring the three coordinate axes .
this implies a first order phase transition ( see d. j. wallace , j. phys .
c * 6 * , 1390 ( 1973 ) ) .
p. bak , s. krinsky , and d. mukamel , phys .
lett . * 36 * , 52 ( 1976 ) .
also : d. mukamel and s. krinsky ; phys .
rev . b * 13 * , 5065 ( 1976 ) , and s. a. brazovskii , i. e. dzyaloshinskii , and b. g. kukharenko , sov .
jetp * 43 * , 1176 ( 1976 ) [ zh . eksp
. teor .
* 70 * , 2257 ( 1976 ) ] .
the general effect of introducing a random field is similar to a reduction in dimensionality .
now , models ( such as the 3-state potts model ) which in @xmath49 have weak first order transitions as an effect of the many component order parameter , tend to become second order in @xmath230 .
this would also lead us to predict a multicritical point ; indeed , such a point is observed in the potts model ( see s. chen , a. m. ferrenberg , and d. p. landau , phys .
lett . * 69 * , 1213 ( 1992 ) , and compare also w. g. wilson , phys .
a * 137 * , 398 ( 1989 ) . ) in principle , one could imagine a intermediate phase which had coexisting antiferromagnetic and spin glass order , which would thus have distinct critical concentrations @xmath41 .
that phase would be analogous to a state of two infinite interpenetrating percolation clusters , which is geometrically possible in @xmath49 . within the rather coarse @xmath40 grid of our simulations , however , only one @xmath41 value is observed along the line @xmath231 .
it is intriguing to speculate on the universality class of the transition as a function of @xmath40 along the line @xmath232 .
the afm order parameter would be expected to behave as in a random
field 3state potts model ( since the effective random fields distinguish among three ordering wavevectors ) yet the spin glass order parameter would behave as at the spin - glass freezing temperature .
our simulations can not address this question ( which is best studied at @xmath233 ) .
this situation is similar to the problem of propagating order in e.g. the 3state potts antiferromagnet on a triangular lattice , where a single chain of bonds is insufficient to force the relationship between two clusters ( see j. adler , r. g. palmer and h. meyer , phys .
* 58 * , 882 ( 1987 ) , and h. fried and m. schick , phys .
b * 41 * , 4389 ( 1990 ) .
however , that problem is unfrustrated , in that the restriction of a pure system ground state to the sites of the diluted system is always _
one _ of the valid ground states ; this is not true for a frustrated system such as the fcc antiferromagnet ( compare ref . ) owing to its frustration and this complicates the propagation of order .
on the basis of simulations , it is argued by r. khn and a. huber , phys .
* 73 * , 2268 ( 1994 ) and by j.k .
kim and a. patrascioiu , phys .
lett . * 72 * , 2785 ( 1994 ) , that the critical exponents of _ unfrustrated _ diluted ising models show a non universal dilution dependence . it is commonly believed that @xmath49 is _ below _ the lower critical dimension for heisenberg spin glasses , but anisotropic spin interactions which are certainly present in dms are believed to drive the transition to the ising class , even though the ordered state may be held together mainly by the isotropic interactions . | in this paper we report on a monte carlo study of a diluted ising antiferromagnet on a fcc lattice .
this is a typical model example of a highly frustrated antiferromagnet , and we ask , whether sufficient random dilution of spins does produce a spin glass phase .
our data strongly indicate the existence of a spin glass transition for spin
concentration @xmath0 : we find a divergent spin glass susceptibility and a divergent spin glass correlation length , whereas the antiferromagnetic correlation length saturates in this regime .
furthermore , we find a first order phase transition to an antiferromagnet for @xmath1 , which becomes continuous in the range @xmath2 .
finite size scaling is employed to obtain critical exponents .
we compare our results with experimental systems as diluted frustrated antiferromagnets as @xmath3 . |
for more than a decade , the gdh and a2 collaborations have been carrying out intense experimental studies aimed at the verification of the famous gerasimov - drell - hearn ( gdh ) sum rule @xcite . a comprehensive overview concerning the status of these experiments
is given in ref .
@xcite .
the gdh sum rule gives an integral relation between anomalous magnetic moment @xmath7 of a spin-@xmath8 particle and the difference of the total photoabsorption cross sections with parallel , @xmath9 , and antiparallel , @xmath10 , photon and target spin alignments and reads [ gdh ] s= ^_0 de _ , where @xmath11 is the photon laboratory energy , @xmath12 is the particle mass , and @xmath13
. equation ( [ gdh ] ) shows that ground - state properties given by @xmath7 and @xmath12 are related to the energy - weighted excitation spectrum of the particle .
the gdh sum rule relies on the basic physics principles of lorenz and gauge invariance , unitarity , crossing symmetry , and an unsubtracted dispersion relation applied to the forward compton amplitude .
therefore , measurements of the right - hand side ( rhs ) of this equation can serve as a fundamental cross - check of these principles . the left - hand side ( lhs ) of eq .
( [ gdh ] ) for the proton is @xmath14 @xmath6b .
the integrand in the rhs of eq .
( [ gdh ] ) at energies from 0.2 to 2.9 gev was determined by the gdh and a2 collaborations in a series of experiments with circularly polarized photons and longitudinally polarized protons @xcite .
involving theoretical predictions from threshold to 0.2 gev and above 2.9 gev , the gdh group obtained for the rhs of eq .
( [ gdh ] ) @xmath15 @xmath6b , which is in good agreement with the gdh sum rule prediction .
the check of the isospin structure of the gdh sum rule requires measurements on the neutron .
however , because of the absence of a stable , dense , free - neutron target , direct measurements for the neutron are impossible so that information on the neutron can be extracted from deuteron data .
experiments with circularly polarized photons and longitudinally polarized deuterons were also performed by the gdh and a2 collaborations @xcite as well as by the legs collaboration @xcite .
in fact , however , a contribution to the neutron gdh integral from deuteron data was extracted only in ref .
@xcite and therewith in the limited energy range of 8151825 mev .
the extraction procedure relied mainly on an assumption that at these energies the incoherent , quasi - free meson production reactions dominate ; i.e. if one supposes that the equality @xmath16 ( @xmath17 ) is valid .
such an assumption seems to be in general agreement with the results obtained by arenhvel , fix , and schwamb ( afs ) @xcite . using theoretical predictions for the unmeasured energy regions , the authors of ref .
@xcite obtained the value of @xmath18 @xmath6b for the rhs of eq .
( [ gdh ] ) which is in agreement with the gdh sum rule value of @xmath19 @xmath6b .
of course , this agreement should not be considered as an experimental verification of the gdh sum rule for the neutron , because the net experimental contribution to this value is only about @xmath20 .
the helicity - dependent cross section on the deuteron in the energy range @xmath21 mev was also measured in ref .
@xcite . combining the obtained data with those from ref .
@xcite and using again the assumption on the incoherency of the proton and neutron contributions in the range @xmath22 mev , the authors of ref .
@xcite obtained the value of @xmath23 @xmath6b in this range .
the said assumption seems to be rather rough below , say , 500 mev , where the @xmath24 and the @xmath25 reactions contribute to the polarized total cross section , nevertheless , the extracted value indicates that the gdh integrals for the proton and neutron are of the same order of magnitude .
together with the verification of the gdh sum rule for the neutron , the helicity - dependent cross sections on the deuteron are also required to test this sum rule for the deuteron itself .
the anomalous magnetic moment of the deuteron is very small ( @xmath26 nm ) because , first , the anomalous magnetic moments of the proton and neutron are almost equal in magnitude but opposite in sign and , second , nucleon spins in the deuteron are predominantly parallel . in view of this , the lhs of eq .
( [ gdh ] ) for the deuteron is @xmath27 @xmath6b , which is more than two orders smaller than the nucleon values .
therefore , there should be large negative contributions almost totally canceling the nucleon contribution .
this point was discussed for the first time in ref .
it was found that a huge negative contribution amounting to about @xmath28 @xmath6b indeed exists and stems from the deuteron photodisintegration reaction .
later on , a comprehensive analysis of different channels in the deuteron gdh integral was performed within the afs framework @xcite .
they evaluated effects from deuteron photodisintegration and single and double - pion production as well as from single-@xmath29 production with the total result of @xmath30 @xmath6b .
if one intends to extract the neutron value for the gdh integral from deuteron data or verify the deuteron gdh sum rule , in either case one has to be confident that theoretical models used in analyses of these data provide reliable descriptions of the helicity - dependent cross sections for different reactions on the deuteron .
as was found in refs .
@xcite , a large contribution to the deuteron gdh integral stems from incoherent single - pion production @xmath4 .
intense experimental investigations of the reaction @xmath31 have recently been performed by the gdh and a2 collaborations @xcite .
references @xcite have provided data on the helicity - dependent total inclusive and semiexclusive cross sections in the energy range from 200 to 800 mev . in the very recent work of the same collaborations @xcite , the helicity dependence of the differential cross section in the @xmath31 channel has been measured for the first time in the @xmath32 region .
there are also measurements of the helicity - dependent total cross section of the reaction @xmath33
@xcite , where @xmath34 ( the incoherent channel ) or @xmath35 ( the coherent channel ) .
the main goal of the present work is to analyze the data of refs .
@xcite on the @xmath31 reaction in the framework of the diagrammatic model built in ref .
@xcite . in that work , we restricted ourselves to the discussion of unpolarized and single spin - dependent observables due to the lack of data on double spin - dependent observables at the time when the work was being prepared .
it was found that the model @xcite provided a quite satisfactory description of all available data except for the unpolarized differential and total cross sections in the @xmath36 channel for which the data were notably overestimated . in the present work ,
we compare predictions of the approach @xcite to the polarized data from refs .
@xcite with special emphasis on the model dependence of our results .
sources for theoretical uncertainties are discussed in ref .
@xcite , to which the interested reader is referred . another principal goal of this work is to evaluate the contribution from the @xmath4 channel to the deuteron gdh integral .
this task requires an extension of the model up to about 1.52 gev .
although in ref .
@xcite we restricted our consideration to the @xmath32 resonance region , the model is applicable at higher energies , too . as was shown in ref .
@xcite , one can expect that in the framework of the diagrammatic approach , contributions due to @xmath1 and @xmath2 final state interaction ( fsi ) are reliably calculated up to a photon energy of 800 mev .
taking into account that these contributions are small above 500 mev , they can obviously be disregarded for @xmath37 mev .
hence , the upper limit for the applicability of the model depends on that for the impulse approximation ( ia ) which , in turn , depends on the applicability domain of an employed elementary operator @xmath3 of pion production on the nucleon . in the framework of the approach @xcite
, this operator is taken from multipole analyses said @xcite and maid @xcite and can be used up to 2 gev for the first analysis and up to 1.65 gev for the second one . the contribution from the @xmath4 channel to the gdh integral up to @xmath38
gev was evaluated in the afs approach @xcite , but the sensitivity of obtained results to the @xmath3 operator was not investigated in that work .
an analysis of ref .
@xcite shows that the value of the integral at integration up to 350 mev manifests its considerable dependence on the elementary operator . because the energy domain from threshold to 350 mev provides only about a half of the total contribution of this reaction to the deuteron gdh integral , it is of high interest to recognize how sensitive the total integral is to the @xmath3 operator . in the present paper ,
we evaluate the integral up to a photon energy of @xmath39 gev and investigate a question of possible theoretical uncertainties for its value .
the work is organized as follows . in sec .
[ kinema ] , the kinematic relations used in the calculations as well as definitions for observables are reviewed .
a brief description of the theoretical model and its ingredients is given in sec .
[ theory ] .
section [ results ] contains results for the unpolarized and helicity - dependent yields and cross sections and their comparison with the data available in the considered kinematic region .
results of the evaluation of the @xmath4 channel contribution to the deuteron gdh integral are also given in sec .
[ results ] . in sec .
[ conclusion ] , we summarize the main conclusions and results .
actual calculations of the reaction amplitude and observables are done in the laboratory frame where @xmath40 , and @xmath41 are the four - momenta of the initial photon and deuteron and the final pion and nucleons , respectively . for the case when the final pion is detected , we take as independent kinematic variables the photon energy @xmath11 , the value of the pion momentum @xmath42 , the pion polar angle @xmath43 , and the solid angle @xmath44 ( @xmath45 , @xmath46 ) of the relative momentum @xmath47 of the final nucleon - nucleon pair .
these variables totally determine the kinematics .
the unpolarized differential cross section is given by [ dcs0 ] = f_nn 16 _
m_2m_1m_d | m_2m_1| t| m_d |^2 , where @xmath48 is the reaction amplitude , and the phase space factor is [ fnn ] f_nn=1(2)^5 , with @xmath49 and @xmath12 is the nucleon mass .
spin states of the two nucleons and deuteron are @xmath50 , @xmath51 , and @xmath52 , respectively , and they are chosen with respect to the @xmath53 axis , which is defined by the photon momentum @xmath54 .
the symbol @xmath55 stands for the photon helicity .
an extra factor of 1/2 must be included in the rhs of eq .
( [ fnn ] ) in the case of charged pion production
. we will be also interested in double polarized observables for the polarized photon beam and polarized deuteron target .
these are the parallel , @xmath56 , and antiparallel , @xmath57 , cross sections defined as [ dcs_p0 ] & = & f_nn 12 _ m_2m_1 ( |m_2m_1| t|+1 + 1|^2+| m_2m_1| t|-1 - 1|^2 ) , + [ dcs_a0 ] & = & f_nn 12 _ m_2m_1 ( |m_2m_1| t|+1 - 1|^2+| m_2m_1| t|-1 + 1|^2 ) . to obtain the semi - inclusive differential cross sections @xmath58 , @xmath59 , and @xmath60 , the right - hand sides of eqs .
( [ dcs0 ] ) , ( [ dcs_p0 ] ) , and ( [ dcs_a0 ] ) have to be integrated over @xmath61 and the solid angle @xmath62 , i.e. , [ dcs ] & = & _ q^min^q^maxf_nn q^2dq d_p_nn , + [ dcs_p ] & = & _ q^min^q^maxf_nn q^2dq d_p_nn , + [ dcs_a ] & = & _ q^min^q^maxf_nn q^2dq d_p_nn .
the upper and lower integration limits in the laboratory frame are defined by the following relations [ q_max ] q^max = q^max(_)=&&1b , + q^min = q^min(_)=&max\{0,&1b } , [ q_min ] where @xmath63 , @xmath64 , @xmath65 , and @xmath6 is the pion mass . for the neutral channel ,
the emitted proton was registered in the experiment @xcite . in this case
we chose as independent kinematic variables the photon energy @xmath11 , the value of the proton momentum @xmath66 , the proton polar angle @xmath67 , and the solid angle @xmath68 ( @xmath69 , @xmath70 ) of the relative momentum @xmath71 of the final meson - neutron pair .
equations analogous to eqs .
( [ q_max ] ) and ( [ q_min ] ) are now of the form [ p_max ] p^max_p = p^max_p(_p)=&&1b , + p^min_p = p^min_p(_p)=&max \{0,&1b } , [ p_min ] where @xmath72 , @xmath73 , and @xmath74 .
only a fraction of the total interval @xmath75 $ ] was accessed in conditions of the experiment @xcite .
there were registered the protons with the absolute value of the momentum @xmath76 , where p^min_exp = p^min_exp(_p)= [ 210 + 0.01(_p-90^)^2 ] mev / c , so that the differential @xmath77 yields @xmath78 , as they were called in ref .
@xcite , were measured rather than the differential cross sections . in analogy to eqs .
( [ dcs])([dcs_a ] ) they are defined as [ y0 ] & = & _ p^min_exp^p_p^max f_n p_p^2 dp_pd_p_n 16
_ m_pm_nm_d | m_pm_n| t| m_d [ y_p ] & = & _ p^min_exp^p_p^max f_n p_p^2 dp_pd_p_n 12 _ m_pm_n ( |m_pm_n| t|+1 + 1|^2+| m_pm_n| t|-1 - 1|^2 ) , + [ y_a ] & = & _ p^min_exp^p_p^max f_n p_p^2 dp_pd_p_n 12 _ m_pm_n ( |m_pm_n| t|+1 - 1|^2+| m_pm_n| t|-1 + 1|^2 ) , where f_n=1(2)^5 , and @xmath79 .
in this section we briefly outline the formalism used to evaluate the @xmath4 reaction amplitude .
it was described in refs .
@xcite , to which the reader is referred for more details .
the diagrammatic approach is exploited to calculate the @xmath4 reaction amplitude .
because in this work we are not interested in the threshold region , we reduce the set of diagrams in comparison to that considered in ref .
specifically , only contributions from the three diagrams shown in figs .
[ diagrams](a)(c ) to the reaction amplitude have been taken into account .
diagram [ diagrams](a ) is often refereed to as ia .
sometimes it is called the pole diagram .
apart from it , the diagrams [ diagrams](b ) and ( c ) corresponding , respectively , to @xmath1 and @xmath2 rescattering in the final state have been considered in the present calculation .
one remark should be made here .
a detailed analysis of the helicity dependence of the @xmath80 reaction up the @xmath81 resonance is presented in ref .
special attention was given in that work to the near - threshold region .
within the approach @xcite the same set of diagrams as presented in fig .
[ diagrams ] was taken into account .
one can expect such a model to be quite realistic near threshold for the description of the charged channels .
but this is not obviously to be the case for the neutral channel . as was shown in ref .
@xcite , a two - loop diagram as in fig .
[ 2loops](a ) which includes simultaneously @xmath2 rescattering in the intermediate state and @xmath82 rescattering in the final state has to be taken into account at threshold energies . a similar mechanism is known to be very important near threshold for coherent @xmath83 photoproduction on the deuteron @xmath25 @xcite .
the importance of diagram ( a ) in fig .
[ 2loops ] at threshold energies is that it contains , together with the @xmath82 scattering amplitude , a block with charged pion photoproduction from the nucleon .
the dominating threshold electric dipole amplitude @xmath84 for the charged channels is about 20 times larger in absolute numbers than that for the neutral channels . for analogous reasons
we also expect that a two - loop diagram [ 2loops](b ) with @xmath1 rescattering in the intermediate state and @xmath2 rescattering in the final state may also be important near threshold .
therefore , the fact that the model including only three contributions shown in figs . [ diagrams](a)(c ) can be used for the description of the @xmath36 reaction at threshold energies is not obvious , and further study is needed to clarify this point .
there is one more two - loop diagram , shown in fig . [ 2loops](c ) , which includes two blocks with @xmath2 scattering . because the @xmath2 scattering lengths are about two orders smaller than those for @xmath1 scattering , the role of this diagram in the threshold region is expected to be much smaller than that of diagrams [ 2loops](a ) and [ 2loops](b ) .
one can anticipate the effect from the above two - loop diagrams to decrease together with increasing @xmath11 when the mentioned enhancement due to the charge pion exchange disappears .
nevertheless , to be fully confident in the smallness of such contributions , their explicit evaluations are needed .
this task , however , is expected to be very difficult , because as we have learned evaluating diagram [ 2loops](a ) in ref .
@xcite , practical calculations are extremely time consuming especially when the inclusive processes are considered , and , therefore , integrations over momenta of final particles have to be carried out .
note that in the case of the exclusive channels , the kinematic regions exist where the diagrams in figs .
[ diagrams](a)[diagrams](c ) can be suppressed , and other mechanisms , in particular , the two - loop diagrams [ 2loops](a)[2loops](c ) become important and have to be taken into account ( see refs .
@xcite for detailed discussions of this point ) .
let us now write explicit expressions for the amplitudes corresponding to diagrams [ diagrams](a)[diagrams](c ) .
one has for the ia amplitude [ ia ] m_2m_1 |t^ia(p_2,p_1,q;k)| m_d = _ m ^m_d_m_2m(p_2 ) m_1| t_nn_1(p_1,q;k ) | m , where @xmath85 is the deuteron wave function having the form [ psi ] ^m_d_m_2m_1(p)=_l=0,2i^l c^1m_s_12 m_212 m_1c^1m_d_1m_slm_ly_l^m_l ( ) u_l(p ) . in eq .
( [ psi ] ) , @xmath86 are the spherical harmonics , and @xmath87 are the clebsch - gordan coefficients .
the @xmath8 and @xmath88 wave function amplitudes @xmath89 and @xmath90 are taken for the cd - bonn potential @xcite .
one should emphasize that calculations with another realistic version of the nucleon - nucleon potential , namely , with the nijmegen model @xcite , have led to essentially the same results , so we have found no sensitivity of our predictions to the choice of a potential .
note that in eq .
( [ ia ] ) and in the amplitudes that follow , we do not write explicitly those corresponding to the permutation @xmath91 , but they are included in the calculations . the symbol @xmath92 in eq .
( [ ia ] ) stands for the elementary operator of pion photoproduction on the nucleon . as is explained in detail in ref .
@xcite ( see , also ref .
@xcite ) , ambiguities in predictions for observables in the @xmath4 reaction stem mainly from the manner in which @xmath93 is embedded into the deuteron . in all recent calculations
@xcite , this operator is taken in the on - shell form and it is parameterized using either the multipole analyses said @xcite and maid @xcite or the effective lagrangian approach ( ela ) @xcite .
although the models are close in their predictions for observables in the reaction @xmath94 , they are not identical .
it is evident that this difference will manifest itself in results of the evaluation of eq .
( [ ia ] ) . to estimate ambiguities caused by different choices of the on - shell operator @xmath93
, we perform our calculations with the latter taken from the recent multipole analyses said @xcite ( solution sp09k ) and maid07 @xcite .
the on - shell operator @xmath92 depends on four invariant amplitudes @xcite for which there exist different options that are equivalent in the on - shell case .
this equivalence is broken when one or both nucleons are off their mass shells .
just such a situation takes place in eq .
( [ ia ] ) where the nucleon @xmath95 is off its mass shell . in this work ,
we evaluate the operator @xmath93 with two sets of the invariant amplitudes , @xmath96 and @xmath97 , as they are called in ref .
@xcite and where their definition can be found .
at least , this allows one to estimate the possible uncertainties in the predicted observables due to off - shell effects in @xmath93 .
the matrix element corresponding to fig . [ diagrams](b ) reads m_2m_1| t^nn(p_2,p_1,q;k ) | m_d&= & m 1p^2_out - p^2_in+i0 + & & _ m_sm_out , m_2m_1| t_nn(e)|p_in , m_smm_sm| t^ia(p_s , p,q;k)|m_d .
[ nnresc ] here @xmath98 is the ia amplitude given by eq .
( [ ia ] ) .
the relative momenta of the @xmath99 pair after and before scattering are @xmath100 and @xmath101 , respectively , and @xmath102 .
the half - off - shell @xmath1 scattering matrix @xmath103 has been obtained by solving an equation of the lippmann - schwinger type for the cd - bonn potential @xcite .
all states with the total angular momentum @xmath104 have been retained in @xmath103 . in the evaluation of eq .
( [ nnresc ] ) there is another ambiguity due to different prescriptions for the determination of the total invariant energy entering the @xmath105 vertex . as was proposed in refs .
@xcite , we define the latter as if the spectator nucleon @xmath106 was on its mass - shell .
other prescriptions can also be found in the literature .
for instance , in ref .
@xcite the active nucleon @xmath95 is put to be on its mass - shell .
we have evaluated eq .
( [ nnresc ] ) with these two prescriptions and found only at most a 2% variation for all observables considered in the next section .
this means that the above ambiguity can be safely disregarded . the matrix element corresponding to fig .
[ diagrams](c ) with @xmath2 rescattering in the final state is m_2m_1| t^n(p_2,p_1,q;k ) | m_d&= & 1p_2 ^
2-p_s^2+i0 f^2_n(|q__i| ) & & _ m_s , where the asterisk denotes variables in the @xmath107 center - of - mass ( c.m . ) frame , i.e. @xmath108 , where @xmath109 and @xmath110 are the c.m .
initial nucleon and pion momenta , respectively , in the @xmath2 scattering vertex .
analogously , @xmath111 , where @xmath112 and @xmath113 are the c.m .
final nucleon and pion momenta , respectively .
the momenta with the asterisk and without it are related to each other through a boost transformation with the velocity @xmath114 .
we take the spectator nucleon @xmath106 to be on its mass shell , i.e. @xmath115 . the total energy @xmath116 of the @xmath107 pairs is @xmath117 .
the half - off - shell @xmath2 scattering matrix @xmath118 has been obtained by solving the lippmann - schwinger equation for a separable energy - dependent @xmath2 potential built in ref .
we have checked that the results remain essentially the same as the @xmath118 found from a meson - exchange model @xcite .
the factor @xmath119 in eq .
( [ pinresc ] ) is introduced to take into account the off - shell nature of the intermediate meson @xmath120 . in accordance with ref .
@xcite , we choose the form factor @xmath121 of the @xmath122 vertex to be of the monopole form f_n(q)= , with a cutoff parameter @xmath123 mev/_c_. the superscript indices @xmath124 " and @xmath125 " in the @xmath118 amplitude in eq .
( [ pinresc ] ) stand for the neutral , @xmath126 , or charge exchange , @xmath127 , channels , respectively .
the ia amplitudes for @xmath83 photoproduction , @xmath128 , and for charged pion photoproduction , @xmath129 , are given by eq .
( [ ia ] ) .
we begin the discission of the results with the unpolarized yield and cross section .
special emphasize is given to their comparison with the very recent experimental data from refs .
a detailed comparison with older data can be found in ref .
@xcite . as a rule
, we will not compare our predictions with the results of other theoretical approaches @xcite because this has already been done in ref .
@xcite .
results for the unpolarized differential yield ( [ y0 ] ) for the @xmath36 reaction at four selected energies from 301 to 455 mev with a step of about 50 mev are shown in fig .
[ pi0unpol ] .
one can see a sizable effect from fsi , which is mainly due to @xmath82 fsi .
the yield is reproduced very well at the proton angles @xmath130 . however , despite of the strong reduction of the unpolarized differential yield due to fsi , the full model clearly overestimates the data in the peak position .
an analogous feature was found also in ref .
@xcite where after integration over complete acceptance , i.e. with @xmath131 given by eq .
( [ p_min ] ) , we observed the predictions to lie well above data from refs .
similar overestimation , although in some variation of size , is present in all recent calculations @xcite .
it is difficult to find a reasonable explanation for this disagreement in the framework of the present model .
first , reasons for a possible failure of the ia calculation are not apparent .
the contribution of ia to the total cross section stems from the quasi - free @xmath132 process , which is evaluated very reliably .
second , as explained in ref .
@xcite , the @xmath1 fsi contribution comes mainly from the kinematic domains where nucleons @xmath95 and @xmath133 in fig .
[ diagrams](b ) are close to their mass shells so that the on - shell parameterization of the elementary operator @xmath93 can be used in eq .
( [ nnresc ] ) .
because modern parameterizations of @xmath93 and @xmath103 are firmly established , we think that the effect of @xmath1 fsi is well under control .
therefore , the noticeable disagreement with the data can not be explained by an underestimation of the @xmath1 fsi contribution .
we also suppose that @xmath2 fsi is evaluated quite reasonably and it can not be responsible for the disagreement , especially if one takes into account that the @xmath2 fsi effect is very small . as discussed in refs .
@xcite , a possible way to resolve the problem is to take into account the absorption of the produced pions by the pair of nucleons .
rough estimations of the absorption effect performed in ref .
@xcite show that it can comprise more than 10% of the total cross section , thus leading to visible attenuation of the production rate .
model uncertainties of the predictions stemming from the choice of the elementary operator @xmath93 are shown in the shaded areas in fig .
[ pi0unpol ] .
these areas have been generated , first , by results with the maid07 and said ( solution sp09k ) multipole analyses and , second , with two sets of the invariant amplitudes , @xmath96 and @xmath97 , as described in the previous section .
one can see that the model dependence is quite small , and even when taken into account , it does not resolve the problem with the description of the data at @xmath134 .
predictions for the unpolarized differential cross section in the @xmath135 reaction are compared in fig .
[ pimnunpol ] to the data from refs .
@xcite at the same energies as in fig .
[ pi0unpol ] .
the agreement is quite reasonable , although the slight overestimation of the data is seen .
the model dependence of the predictions is mainly due to the choice of the multipole analysis and is noticeable only at the lowest energy .
it is compatible with the data errors of ref .
together with increasing @xmath11 this dependence diminishes and becomes quite small above 350 mev . as is seen in fig .
[ piplunpol ] , the agreement of the predictions and the unpolarized @xmath136 data is almost perfect .
again , the visible dependence of the results on the choice of the multipole analysis occurs only at 301 mev .
taking into account the results for the unpolarized differential cross sections in the charged channels , one could expect the sum of the total cross sections for these two channels to be in reasonable agreement with the data but with a possible slight overestimation because of the above overestimation in the @xmath137 reaction .
however , as is seen in fig .
[ stot_charged ] , the deviation from the data of the gdh and a2 collaborations @xcite at @xmath138 300 mev is more serious . at present
, we are not aware of reasons for this disagreement .
figure [ stot_charged ] shows also that our results are in good agreement with these of afs up to the peak position at about 320 mev , but at higher energies our cross sections overestimate the afs results .
next we discuss the polarized yield and cross section .
figure [ pi0pol ] shows the helicity - dependent yield difference [ dy ] = - for the @xmath77 channel .
similar to the unpolarized yield , one can see a sizable reduction of @xmath139 in ia due to fsi .
however , the striking distinction from the unpolarized case is that now the model reasonably reproduces all the data not only those at @xmath130 .
this might indicate that the above overestimation in the unpolarized yield is due to an overestimation in both @xmath140 and @xmath141 , but it disappears in the difference @xmath139 .
it is seen in fig .
[ pi0pol ] that the dependence of @xmath142 on the elementary operator is weak below 350 mev , but it becomes quite visible at higher energies .
figures [ pimnpol ] and [ piplpol ] show the helicity - dependent differential cross section difference [ dsigma ] = - for the @xmath137 and @xmath136 channels , respectively . there is a good agreement with our results within the uncertainties of the theoretical predictions .
only a few @xmath137 data points are underestimated at @xmath143 .
the above uncertainties are quite significant only in the @xmath136 channel and just in the angular region overlapped in the experiment @xcite .
an integration of eq .
( [ dsigma ] ) over the solid angle @xmath43 gives the difference @xmath144 of the total cross sections .
this difference for the semi - exclusive channels @xmath145 and @xmath146 has been measured in refs .
@xcite at energies from 200 to 430 mev .
because we do not have a model for the coherent reaction @xmath25 , we are able to make predictions for the second channel . our results for the @xmath147 channel are shown in fig .
[ stotpma ] . within theoretical uncertainties ,
the model satisfactory describes the data of ref .
@xcite below 300 mev , but there exists noticeable overestimation at higher energies , i.e. one observes the situation analogous to that for the unpolarized cross section .
reasonable agreement with the data of ref .
@xcite is achieved except for two points in the peak region .
our results are in quite good agreement with the afs predictions but overestimate these from ref .
@xcite at @xmath148 mev . having measured the unpolarized yield and cross section as well as @xmath149 and @xmath150 , the authors of ref .
@xcite have attempted to separate the parallel and antiparallel components of the yield and cross section . to do this
, they used the follows relations : [ ds ] & = & 13 ( 2 + ) , + [ t20 ] t_20&=&3 ( - ) , where @xmath151 and @xmath152 are the components of the cross section corresponding to the deuteron spin states @xmath153 and @xmath154 , respectively , and @xmath155 is the tensor target asymmetry @xcite . note that the component @xmath151 is the average of the parallel and antiparallel cross sections [ ds_paral ] = 12 ( + ) .
it is follows from eqs .
( [ ds ] ) and ( [ t20 ] ) that [ ds_t20 ] = .
evaluations of @xmath155 in refs .
@xcite showed that in the angular regions accessed in conditions of the experimental setup of ref .
@xcite , i.e. at @xmath156 , the absolute value of @xmath155 does not exceed @xmath157 . therefore , one has the following approximate relations for @xmath156 : [ dsparal ] , .
using eqs .
( [ dsigma ] ) , ( [ ds ] ) , ( [ ds_paral ] ) , and ( [ dsparal ] ) one obtains [ dsp / a ] 12 . it should be noted that the semi - inclusive asymmetry @xmath155 resulting from the integration of the corresponding right - hand sides over the kinematic regions appearing in eqs .
( [ y0])([y_a ] ) was not evaluated in refs .
however , we have checked that such an integration gives @xmath158 at @xmath159 and , therefore , the relations similar to eqs .
( [ dsparal ] ) and ( [ dsp / a ] ) are valid for the measured yields as well .
figures [ dy_pa][spipl_pa ] demonstrate our predictions for the parallel and antiparallel yields and cross sections . a comparison with the experimental results from ref .
@xcite is also presented in these figures .
one can see that the description of the data is quite reasonable , but not perfect .
such an agreement could be anticipated for the charged channels , but it is rather unexpected for the @xmath77 reaction .
currently , we are not aware of the reasons why the model that describes reasonably well the @xmath56 and @xmath57 components of the polarized yield , nevertheless , fails in reproducing the unpolarized yield .
we do not discuss here the shape of the angular behavior of the @xmath56 and @xmath57 components in the yield and differential cross section because the corresponding analysis has already been performed in ref .
as shown in that work , this behavior can be interpreted in terms of pion photoproduction on single nucleons , and it is a consequence of the dominance of quasi - free reaction mechanisms .
let us discuss now the contribution of the @xmath4 reaction to the deuteron gdh integral and compare our results with those obtained in refs . .
figure [ gdhint ] presents the difference @xmath160 as a function of the photon energy @xmath11 for three channels . as was shown in refs .
@xcite , the behavior of the difference repeats mainly that for the elementary reaction @xmath3 with some distortion due to nuclear effects .
for instance , the negative value of @xmath160 in the charged channels at threshold energies is because of the dominance of the @xmath84 multipole producing the antiparallel transition .
the @xmath161 multipole provides the strong positive value in the @xmath32 region that fig .
[ gdhint ] clearly demonstrates . as seen in fig .
[ gdhint ] , our results are quite close to those obtained in ref . @xcite
. visible difference may be seen only in the neutral channel in the first resonance region .
the above disagreement also manifests itself in the value of the gdh integral , integrated up to 1.5 gev .
the upper limit of 1.5 gev is chosen in accordance with an option from ref .
@xcite . taking into account the theoretical uncertainties
, we obtain the values of @xmath162 , @xmath163 , and @xmath164 ( in units of @xmath6b used also below ) for the @xmath77 , @xmath137 , and @xmath136 reactions , respectively , to be compared to the afs predictions , respectively , @xmath165 , , @xmath166 , and @xmath167 .
one can see that our value for the @xmath77 channel exceeds that of afs by about @xmath168 .
note here that the use of the said parametrization allows one to extend the upper limit in the gdh integral up to 2 gev .
with such an extension we have not found any visible changes in the result for 1.5 gev .
this means that the @xmath122 term in the deuteron gdh integral essentially reaches convergence at these energies .
the evaluation of the gdh integral , integrated up to 350 mev , has been performed in ref .
results of that work with the maid03 model for the elementary @xmath3 operator are , respectively , @xmath169 , @xmath170 , and @xmath171 for the @xmath77 , @xmath137 , and @xmath136 reactions , which are in complete agreement with our predictions of @xmath172 , @xmath173 , and @xmath174 , respectively . we have also studied the influence of the different parametrizations for the elementary @xmath3 amplitude on the value of the gdh integral , integrated to the maximum energy of 1.65 gev . as is seen in table [ table ] , the theoretical uncertainties in the @xmath77 reaction are quite small .
they are more pronounced in the charged channels , especially in the @xmath136 one .
this considerable dependence of the results is mainly due to the multipole analyses , but the effect of a choice of the invariant amplitudes is also quite visible . note that the noticeable dependence of the nucleon gdh integral evaluated with the said and maid analyses was found in refs .
the sensitivity of the @xmath175 contribution to the deuteron gdh integral , integrated up to 350 mev , to a choice of the elementary operator have been studied in ref .
@xcite , and considerable dependence of the results obtained with the ela and maid03 operators have been found . .
sensitivity of the gdh integral ( in @xmath6b ) , integrated up to 1.65 gev in the @xmath175 reaction to the parametrization of the elementary @xmath3 amplitude . [ cols="<,^,^,^,^",options="header " , ] [ table ] keeping in mind all the numbers given in table [ table ] , we can quote our final value as @xmath5 @xmath6b for the contribution from the @xmath122 channel to the gdh integral for the deuteron , integrated up to 1.65 gev . as seen from table [ table ] , the main part of the uncertainties stems from the @xmath136 channel .
it is worth mentioning that within the uncertainties of @xmath176 @xmath6b , the afs result of @xmath30 @xmath6b @xcite for the total contribution to this integral for the deuteron is compatible with the gdh sum rule prediction of @xmath27 @xmath6b .
the central aim of the present work is the study of the @xmath177 reaction with the polarized photon and deuteron and its contribution to the gdh sum rule for the deuteron . to evaluate the reaction amplitude ,
we have used the diagrammatic model elaborated previously in ref .
@xcite , which takes into account diagrams corresponding to ia as well as @xmath1 and @xmath2 interactions in the final state . for the elementary operator of pion photoproduction on the nucleon ,
its on - shell form generated by the recent multipole analyses , said and maid , has been used .
particular emphasis has been placed on the discussion of possible uncertainties introduced into the model by this operator .
comparing our results with the recent experimental data obtained by the gdh and a2 collaborations in the energy domain from 300 to 450 mev @xcite , we have rediscovered a known problem consisting in the sizable overestimation of the unpolarized data for the neutral channel . at the same time , the unpolarized data for the charged channels are reasonably reproduced .
we have found that the helicity - dependent differential yield difference @xmath139 for the neutral channel and differential cross section difference @xmath178 for the charged channels are reproduced fairly well .
also , the model provides a quite satisfactory description of the separate parallel and antiparallel yields and cross sections in all the channels .
we have also studied the uncertainties arising from the form of the elementary operator of pion production on the nucleon . regarding the unpolarized observables ,
the sizable dependence has been found only for the charged channels at the lowest energy of about 300 mev .
the difference @xmath178 manifests the weak sensitivity to the operator in the @xmath137 channel in the full energy domain , and the difference @xmath139 in the @xmath77 channel does the same below 350 mev .
however , a large sensitivity has been found in the @xmath136 reaction from 300 to 450 mev at @xmath179 . within our model
, we have evaluated the contribution of the @xmath122 channel to the gdh integral for the deuteron , integrated up to 1.65 gev .
in addition , the influence of the elementary operator on its value has been estimated .
our final result is @xmath5 @xmath6b .
the uncertainty of @xmath176 @xmath6b originates mainly from the @xmath136 channel . accepting these values as error bars for the value of @xmath30 @xmath6b obtained in the afs framework @xcite for the deuteron gdh integral
, one can conclude that within these error bars , the afs result is compatible with the gdh sum rule prediction of @xmath27 @xmath6b .
further improvements of the present model are needed to resolve , at last , the problem with the strong overestimation of the unpolarized cross section in the @xmath77 channel .
in particular , one should evaluate contributions of the two - loop diagrams which take into account effectively the absorption of the photoproduced @xmath180 meson by the @xmath1 pair .
on the experimental side , it would be very important to continue polarization measurements in the @xmath4 channel .
specifically , data on the tensor asymmetry @xmath155 together with these on the unpolarized cross section could provide the currently absent information on the @xmath154 component of the reaction amplitude .
r. a. arndt , w. j. briscoe , i. i. strakovsky , and r. l. workman , phys .
c * 66 * , 055213 ( 2002 ) .
m. dugger _ et al .
c * 76 * , 025211 ( 2007 ) .
m. dugger _ et al .
c * 79 * , 065206 ( 2009 ) . | helicity - dependent incoherent pion photoproduction in the reaction @xmath0 is studied in the framework of the diagrammatic approach .
contributions to the reaction amplitude from diagrams corresponding to impulse approximation as well as @xmath1 and @xmath2 interactions in the final state have been evaluated .
the elementary @xmath3 operator is taken from the maid and said models . a detailed comparison of the predictions with recent experimental data by the gdh and a2 collaborations at energies below 500 mev is presented . reasonable agreement with the data on the yields and
cross sections for polarized beam and polarized target has been achieved in all channels .
the unpolarized data of the gdh and a2 collaborations have also been analyzed within the approach .
a strong overestimation for the neutral channel has been found . at the same time
, the model provides a quite satisfactory description of the unpolarized data for the charged channels .
the sensitivity of the obtained results to the choice of the elementary @xmath3 operator is discussed in detail .
the contribution of the @xmath4 reaction to the gdh sum rule for the deuteron up to a photon energy of 1.65 gev has been evaluated with the result of @xmath5 @xmath6b . |
[ sec : general ] a _ cylinder _ for a translation surface is a topological annulus which is isometric to @xmath37 , and is not properly contained in a larger cylinder . here
@xmath38 and @xmath39 are respectively the _ height _ and _ circumference _ of the cylinder . the image of @xmath40 under such an isometry
is called a _ waist curve . _ a cylinder
is always bounded by saddle connections parallel to the waist curve .
a cylinder is called _ simple _ if both of its boundary consists of a single saddle connection . [
prop : central symmetry ] suppose @xmath41 is a translation surface of genus at least two , which has a convex presentation as a @xmath42-gon
. then @xmath41 contains a simple cylinder @xmath43 , and there is a direction @xmath44 transverse to the circumference of @xmath43 , such that @xmath44 is the direction of a saddle connection contained in @xmath43 passing between its boundary edges ( i.e. a ` diagonal ' ) , and any path in @xmath41 in direction @xmath44 either lies in a saddle connection or intersects both boundary components of @xmath43 .
furthermore , if the @xmath42-gon is strictly convex , then there is a hyperelliptic involution of @xmath41 fixing @xmath43 , all paths in @xmath41 in direction @xmath44 intersect any boundary component of @xmath45 of @xmath43 , and the return map @xmath46 obtained by moving in direction @xmath44 is an interval exchange on @xmath47 intervals , corresponding to the ` inversive permutation ' @xmath48 .
denote the convex @xmath42-gon representing @xmath41 by @xmath49 . since @xmath41 has genus at least two , @xmath50
let @xmath51 denote the edges of @xmath49 , in cyclic counterclockwise order .
let @xmath52 denote the angle that the edge @xmath53 makes with the horizontal direction where the @xmath53 are oriented so that @xmath49 is on their left .
we consider @xmath52 modulo @xmath54 and write @xmath55 if the short path along the boundary of the circle of directions from @xmath56 to @xmath44 is in the counterclockwise direction .
then convexity ( strict convexity ) is equivalent to @xmath57 ( resp .
@xmath58 ) for all @xmath59 . for each @xmath59
we denote by @xmath60 the index for which @xmath53 and @xmath61 are identified in @xmath41 .
the convex hull of @xmath62 and @xmath63 is a parallelogram in the plane .
denote its sides @xmath64 in the usual order .
this parallelogram maps to a cylinder @xmath43 in @xmath41 , which is simple if and only if the sides @xmath65 and @xmath66 contain no vertices of @xmath49 in their interior .
it is clear from convexity that any path in @xmath41 in direction @xmath67 intersects the boundary edges of @xmath43 , so to prove the first assertion we have only to find @xmath68 for which @xmath43 is simple .
suppose @xmath43 is not simple and suppose with no loss of generality that @xmath65 contains vertices in its interior , then @xmath69 and @xmath70 assume first that @xmath71 , that is an edge of @xmath49 has three consecutive parallel edges . then replacing @xmath68 with @xmath72 ,
we see from that neither @xmath73 nor @xmath74 can contain vertices in their interior , i.e. the parallelogram @xmath75 maps to a simple cylinder in @xmath41 , as required .
now assume @xmath76 .
then the equations @xmath77 imply that the angles @xmath52 assume only the four values @xmath78 , i.e. @xmath49 is a parallelogram .
if it has three consecutive parallel edges we are done as in the previous case . otherwise @xmath79 and in each direction there are two consecutive parallel edges . since there are no spurious vertices on @xmath49 , in this case the gluing must be @xmath80 , and one sees that for any @xmath68 , @xmath81 maps to a simple cylinder , as required .
now suppose @xmath49 is strictly convex . we claim that @xmath82 indeed , if this is not true then ( replacing @xmath68 with @xmath83 and re - indexing cyclically ) we would have for some @xmath84 : @xmath85 on the other hand , since @xmath84 there is some @xmath86 so that both @xmath86 and @xmath87 are in @xmath88 .
this implies @xmath89 contradicting .
the remaining assertions of the proposition follow easily from , where the involution is the symmetry with respect to the center of the parallelogram representing the cylinder @xmath43 .
the following is an immediate consequence : [ cor : hyperelliptic stratum ] suppose that @xmath41 is a translation surface with a presentation as a strictly convex @xmath42-gon , with @xmath50 .
then : * all vertices of the @xmath42-gon are singularities for the translation structure . *
if @xmath7 is even then @xmath41 belongs to the hyperelliptic connected component of the stratum @xmath10 where @xmath90 is the genus of @xmath41 . *
if @xmath7 is odd then @xmath41 belongs to the hyperelliptic connected component of the stratum @xmath11 where @xmath91 is the genus of @xmath41 .
it was proved by kontsevich and zorich @xcite that in every connected component of every stratum there is a dense set of surfaces with a one - cylinder presentation .
in particular having a convex presentation places no constraint on the topology of the surface .
corollary [ cor : hyperelliptic stratum ] shows that having a strictly convex presentation does entail a topological restriction .
we will need another consequence of proposition [ prop : central symmetry ] .
[ cor : short saddle connection ] suppose @xmath41 is a translation surface of genus @xmath29 represented by a strictly convex polygon @xmath49 with opposite sides identified , and suppose @xmath43 is a simple cylinder in @xmath41 joining opposite sides of @xmath49 .
let @xmath92 be the saddle connection joining the top and bottom of @xmath43 which is represented by two opposite segments on the boundary of @xmath49 .
then any saddle connection parallel to @xmath92 is longer than @xmath92 , and @xmath41 has no other simple cylinder which is parallel to @xmath43 and contains a diagonal parallel to @xmath92 .
in particular , a surface with a strictly convex presentation does not have a nontrivial translation automorphism .
assume with no loss of generality that @xmath43 is represented by a central square in @xmath49 with opposing vertical edges identified .
since @xmath93 does not cover the entire surface @xmath41 , and since @xmath49 is strictly convex , any vertical saddle connection in @xmath49 must pass through @xmath43 at least once and have endpoints disjoint from the top and bottom edges of @xmath43 ; in particular its length is greater than the height of @xmath43 .
thus , if @xmath94 is another cylinder in @xmath41 which is parallel to @xmath43 , with a diagonal @xmath95 parallel to @xmath92 , the interior of @xmath95 must intersect @xmath43 , that is the interiors of @xmath43 and @xmath94 intersect .
this is not possible for parallel cylinders .
this proves the first assertion , and the second assertion follows by considering the image of @xmath43 under a translation automorphism .
the following simple observation , noted by veech @xcite , is at the heart of our discussion .
its proof is immediate from the definitions and is left as an exercise to the reader .
[ prop : g invariant closed ] let @xmath96 be a connected component of a stratum of translation surface .
then the subsets @xmath97 of @xmath96 consisting respectively of surfaces with no convex ( resp .
, no strictly convex ) presentations are @xmath13-invariant , and @xmath98 is closed .
[ sec : examples ] in this section we give four infinite lists of examples of surfaces which do not admit convex presentations . these will all be _ lattice surfaces _ , i.e. surfaces @xmath41 for which the @xmath13-orbit @xmath99 is closed .
equivalently , their _ veech group _
@xmath100 is a lattice in @xmath13 .
as veech showed in @xcite , on a lattice surface , the direction of a core curve of a cylinder is always _ completely periodic _ , all non - critical leaves in that direction are closed .
moreover the @xmath101-orbits of completely periodic directions are in bijective correspondence with the cusps of the quotient @xmath102 , i.e. with conjugacy classes of maximal parabolic subgroups of @xmath101 . in proving that a lattice surface has no convex presentation , our strategy will be to examine the various completely periodic directions on @xmath41 , and show that they can not be the direction of a core curve of a simple cylinder as in proposition [ prop : central symmetry ] .
we start with the _ veech double @xmath7-gon _ , obtained by taking @xmath103 and forming two regular @xmath7-gons @xmath104 , where @xmath105 is the reflection of @xmath106 in ( any ) one of its side , and gluing each edge of @xmath106 with the parallel edge of @xmath105 .
see figure [ fig : double n gon ] . in 1.7 cm ( 0,-/2 ) + + ( -18: ) + + ( 54: ) + + ( 126: )
+ + ( 198: ) + + ( -18:- ) + + ( 54:- ) + + ( 126:- )
+ + ( 198:- ) cycle ; ( 0,-/2 ) + ( 90: ) ; ( 0,-/2 ) node [ inner sep=2pt , circle , draw , fill = white ] + + ( -18: ) node [ inner sep=2pt , circle , draw , fill = white ] + + ( 54: ) node [ inner sep=2pt , circle , draw , fill = white ] + + ( 126: ) node [ inner sep=2pt , circle , draw , fill = white ] + + ( 198: ) node [ inner sep=2pt , circle , draw , fill = white ] + + ( -18:- ) node [ inner sep=2pt , circle , draw , fill = white ] + + ( 54:- ) node [ inner sep=2pt , circle , draw , fill = white ] + + ( 126:- ) node [ inner sep=2pt , circle , draw , fill = white ] ; in 1.5 cm ( 0,-/2 ) + + ( -30: ) + + ( 30: ) + +
( 90: ) + + ( 150: ) + + ( 210: ) + + ( -30:- ) + + ( 30:- ) + + ( 90:- ) + + ( 150:- ) + + ( 210:- ) cycle ; ( 0,-/2 ) + ( 90: ) ; ( 0,-/2 ) node [ inner sep=2pt , circle , draw , fill = white ] + + ( -30: ) node [ inner sep=2pt , circle , fill = black ] + + ( 30: ) node [ inner sep=2pt , circle , draw , fill = white ] + + ( 90: ) node [ inner sep=2pt , circle , fill = black ] + + ( 150: ) node [ inner sep=2pt , circle , draw , fill = white ] + + ( 210: ) node [ inner sep=2pt , circle , fill = black ] + + ( -30:- ) node [ inner sep=2pt , circle , draw , fill = white ] + + ( 30:- ) node [ inner sep=2pt , circle , fill = black ] + + ( 90:- ) node [ inner sep=2pt , circle , draw , fill = white ] + + ( 150:- ) node [ inner sep=2pt , circle , fill = black ] ; [ prop : regular polygons ] the veech double @xmath7-gons have no convex presentations .
veech @xcite proved this assertion for @xmath7 odd and asserted its validity for all @xmath7 , without proof .
our argument is different from that of veech .
veech @xcite showed that the double @xmath7-gon surfaces are lattice surfaces , and computed their veech group , showing that it has either one cusp ( when @xmath7 is odd ) or two ( when @xmath7 is even ) . if @xmath7 is odd , in the cylinder decomposition corresponding to the unique cusp , the only simple cylinder lies at the left and right sides of the double @xmath7-gon ( vertical direction in figure [ fig : double n gon ] ) and this cylinder does not intersect all leaves in the direction of the corresponding diagonal .
thus the necessary condition of proposition [ prop : central symmetry ] is not satisfied . if @xmath7 is even then one cylinder decomposition has no simple cylinders at all , and the other has one simple cylinder , treated as in the previous case of @xmath7 odd .
now let @xmath107 . the _ escher staircase _
is obtained by a cyclic gluing of squares as in figure [ fig : escher ] .
that is , if the squares are labelled @xmath108 , then square @xmath109 is attached to square @xmath110 along its horizontal sides , and to square @xmath111 along its two vertical sides ( all calculations modulo @xmath42 ) .
( 0,0 ) rectangle ( 6,6 ) ; ( 0,0 ) node [ above ] @xmath112 + + ( 2,0 ) + + ( 2,0 ) + + ( 0,2 ) + + ( 2,0 ) + + ( 0,2 ) node [ above ] @xmath113 + + ( 2,0 ) + + ( 0,2 ) node [ above ] @xmath112 + + ( -2,0 ) + + ( -2,0 ) + + ( 0,-2 ) + + ( -2,0 ) + + ( 0,-2 ) node [ above ] @xmath113 + + ( -2,0 )
+ + ( 0,-2 ) cycle ; * x*in 0,2,4,6 ( * x*,*x * ) circle ( 2.2pt ) ; ( * x*+2,*x * ) circle ( 2pt ) ; ; * x*in 0,2,4 ( * x*+4,*x * ) circle ( 2.2pt ) ; ( * x*,*x*+2 ) circle ( 2pt ) ; ; ( 0,0 ) ( 1.00000 , 0.00000 ) ( 0.67844 , 0.15485 ) ; ( 0,0 )
( 0.90096 , 0.43388 )
( 0.54407 , 0.43388 ) ; ( 0,0 )
( 0.62348 , 0.78183 )
( 0.30193 , 0.62698 ) ; ( 0,0 )
( 0.22252 , 0.97492 )
( 0.00000 , 0.69589 ) ; ( 0,0 )
( -0.22252 , 0.97492 ) ( -0.30193 , 0.62698 ) ; ( 0,0 )
( -0.62348 , 0.78183 ) ( -0.54407 , 0.43388 ) ; ( 0,0 )
( -0.90096 , 0.43388 ) ( -0.67844 , 0.15485 ) ; ( 0,0 ) ( -1.00000 , 0.00000 ) ( -0.67844 , -0.15485 ) ; ( 0,0 )
( -0.90096 , -0.43388 ) ( -0.54407 , -0.43388 ) ; ( 0,0 ) ( -0.62348 , -0.78183 ) ( -0.30193 , -0.62698 ) ; ( 0,0 )
( -0.22252 , -0.97492 )
( 0.00000 , -0.69589 ) ; ( 0,0 )
( 0.22252 , -0.97492 )
( 0.30193 , -0.62698 ) ; ( 0,0 )
( 0.62348 , -0.78183 )
( 0.54407 , -0.43388 ) ; ( 0,0 )
( 0.90096 , -0.43388 )
( 0.67844 , -0.15485 ) ; ( 1.00000 , 0.00000 )
( 0.67844 , 0.15485 )
( 0.90096 , 0.43388 )
( 0.54407 , 0.43388 )
( 0.62348 , 0.78183 )
( 0.30193 , 0.62698 ) ( 0.22252 , 0.97492 )
( 0.00000 , 0.69589 )
( -0.22252 , 0.97492 )
( -0.30193 , 0.62698 ) ( -0.62348 , 0.78183 ) ( -0.54407 , 0.43388 )
( -0.90096 , 0.43388 ) ( -0.67844 , 0.15485 ) ( -1.00000 , 0.00000 ) ( -0.67844 , -0.15485 ) ( -0.90096 , -0.43388 )
( -0.54407 , -0.43388 ) ( -0.62348 , -0.78183 )
( -0.30193 , -0.62698 ) ( -0.22252 , -0.97492 )
( 0.00000 , -0.69589 )
( 0.22252 , -0.97492 )
( 0.30193 , -0.62698 ) ( 0.62348 , -0.78183 )
( 0.54407 , -0.43388 )
( 0.90096 , -0.43388 )
( 0.67844 , -0.15485 ) cycle ; [ fig : escher ] [ prop : escher ] the escher staircases have no convex presentations .
one shows using schmithsen s algorithm @xcite that there are two cusps , corresponding to the horizontal and unit slope directions in figure [ fig : escher ] .
these directions have no simple cylinders .
ward @xcite constructed translation surfaces by applying the ` unfolding procedure ' to the billiards in the @xmath114 triangles , and proved that the corresponding surfaces are lattice surfaces . see figure [ fig : escher ] .
[ prop : ward ] the ward examples have no convex presentations .
as proved by ward , in his examples the veech group has one cusp .
therefore the horizontal direction in figure [ fig : escher ] represents the unique completely periodic direction ( up to the @xmath101-action ) .
to locate the cylinders in this direction one simply draws all horizontal segments starting at vertices of the polygon in the figure ; this analysis reveals that there are no simple cylinders in the horizontal direction .
therefore the ward surfaces have no simple cylinders at all .
bouw and mller @xcite found an infinite sequence of lattice surfaces @xmath115 .
[ prop : bm ] when @xmath116 and @xmath117 are not both even , the surface @xmath115 has no convex presentation .
hooper @xcite described the bouw - mller examples in terms of a decomposition into semi - regular polygons ; see also @xcite , where it is shown that hooper s surfaces in fact coincide with @xmath115 .
the work of @xcite also showed that the veech group of @xmath115 is either the @xmath118 triangle group ( when @xmath119 ) , or the @xmath120-triangle group ( when @xmath121 are odd ) , and hence these examples have a unique ( up to the veech group action ) cylinder decomposition pattern .
hence any simple cylinder , if it existed , would have to be horizontal in a direction of a semi - regular polygon decomposition .
it can be checked from hooper s description ( see figure [ fig : bm ] ) that when @xmath116 and @xmath122 are not both even , the surfaces @xmath115 have no simple cylinders . .
each edge is identified with an edge in a neighboring polygon . ]
the property of having no simple cylinder is also @xmath13-invariant and closed
. it would be interesting to classify all surfaces without simple cylinders .
a similar closed and @xmath13-invariant property is having no _ semisimple cylinder _ ,
i.e. a cylinder which is bounded on at least one side by a single saddle connection .
[ subsec : square zoo ] we conclude this zoological survey with four interesting square - tiled surfaces .
no menagerie is complete without the legendary _ eierlegende wollmilchsau _ and _ ornithorhynque _ ( see @xcite ) .
one of the many surprising features of these notoriously eccentric square tiled surfaces , is the lack of simple cylinders .
in particular they have no convex presentations . [
prop : interesting square tiled ] the surface made of four squares ( see the left hand side of figure [ fig : lots of squares ] ) has no convex presentation .
any square - tiled surface in @xmath1 made of seven squares or fewer ( see the two surfaces on the right hand side of figure [ fig : lots of squares ] ) has no strictly convex presentations .
= [ draw , inner sep=1pt , circle ] ( 1,0 ) ( 1,1 )
( 2,2 ) ; ( 0,0 ) node [ circle , semithick , fill = white ] node ( 1,0 ) node [ circle , very thin , fill = black ] node ( 2,0 ) node [ circle , semithick , fill = white ] node ( 2,1 ) node [ circle , semithick , fill = white ] node ( 3,1 ) node [ circle , very thin , fill = black ] node ( 3,2 ) node [ circle , semithick , fill = white ] node ( 2,2 ) node [ circle , very thin , fill = black ] node ( 1,2 ) node [ circle , semithick , fill = white ] node ( 1,1 ) node [ circle , very thin , fill = black ] node ( 0,1 ) node [ circle , semithick , fill = white ] node ( 0,0 ) cycle ; ( 0,0 ) node @xmath123 node@xmath124 ( 2,0 ) node
node@xmath125 ( 4,0 ) node @xmath123 node@xmath126 ( 6,0 ) node
node@xmath127 ( 7,0 ) node @xmath123 node@xmath128 ( 7,1 ) node node@xmath124 ( 5,1 ) node @xmath123 node@xmath125 ( 3,1 ) node node@xmath126 ( 1,1 )
node @xmath123 node@xmath127 ( 0,1 ) node node@xmath128 ( 0,0 ) cycle ; ( 1,0 )
( 1,1 ) ( 2,0 )
( 2,1 ) ( 3,0 )
( 3,1 ) ( 4,0 ) ( 4,1 ) ( 5,0 )
( 5,1 ) ( 6,0 )
( 6,1 ) ; ( 0,0 ) node @xmath123 node@xmath129 ( 3,0 ) node node@xmath130 ( 5,0 ) node @xmath123 node@xmath131 ( 6,0 ) node node@xmath132 ( 7,0 ) node @xmath123 node@xmath133 ( 7,1 ) node node@xmath129 ( 4,1 ) node @xmath123 node@xmath130 ( 2,1 ) node
node@xmath131 ( 1,1 ) node @xmath123 node@xmath132 ( 0,1 ) node
node@xmath133 ( 0,0 ) cycle ; ( 1,0 )
( 1,1 ) ( 2,0 ) ( 2,1 ) ( 3,0 )
( 3,1 ) ( 4,0 ) ( 4,1 ) ( 5,0 )
( 5,1 ) ( 6,0 )
( 6,1 ) ; [ fig : lots of squares ] for the first assertion , checking the @xmath134-orbit ( see @xcite )
one finds that the @xmath13-orbit of this surface has two cusps , none of whose corresponding cylinder decompositions is a one - cylinder decomposition .
only one of the cusps has a simple cylinder @xmath43 , and one can verify that the foliation in the direction of any diagonal of @xmath43 has leaves which do not cross @xmath43 and are not on saddle connections .
so by proposition [ prop : central symmetry ] , the surface has no convex presentation . for the second assertion ,
let @xmath135 be square - tiled , and suppose @xmath43 is a simple cylinder of a convex presentation , as in proposition [ prop : central symmetry ] .
we claim that @xmath43 has area at least four ( where we normalize area so that each square is of unit area ) . indeed , replacing @xmath41 by its image under an element of @xmath134
we can represent @xmath43 as a rectangle in the plane with corners at integer points , where all vertices of @xmath41 remain at integer points , such that the two boundary edges of @xmath43 are horizontal .
in this situation , a generalized diagonal is an edge crossing @xmath43 from top to bottom , where it is allowed to wrap around the circumference of @xmath43 . for rectangles of area less than four , straight lines starting at integer points in the plane , in the directions of generalized diagonals , can intersect the boundary of @xmath43 in at most three points . however according to proposition [ prop : central symmetry ] ,
there must be four intersection points , corresponding to the discontinuity points of the interval exchange .
this proves the claim .
now let @xmath49 be a convex polygon representing @xmath41 .
since @xmath136 , @xmath49 is a decagon , and @xmath137 is a union of two quadrilaterals . identifying their edges which are boundary components of @xmath43
, we obtain a new strictly convex polygon @xmath138 .
the vertices of @xmath138 are integer points and @xmath138 is an octagon of area at most three . recall pick s formula @xcite , which says that the area of @xmath138 is @xmath139 , where @xmath59 is the number of interior integer points and @xmath113 is the number of integer points along the edges . by construction @xmath140 , so in fact @xmath141 , @xmath142 and the area of @xmath138 is exactly three .
label the sides of @xmath138 by vectors @xmath143 , cyclically in the counterclockwise direction .
applying @xmath134 again we can assume that @xmath144 meet at the origin and lie respectively along the @xmath145-axis and @xmath146-axis . since the eight vertices are the only integer points on sides , @xmath144 are of unit length . by strict convexity ,
the internal angle between @xmath147 and @xmath148 is more than @xmath149 and so is the internal angle between @xmath150 and @xmath151 .
this forces @xmath152 to be an interior point of @xmath138 , a contradiction .
[ sec : genus 2 ] in this section we summarize results about genus two surfaces which will be crucial for our discussion .
these results incorporate work of many authors , notably mcmullen .
we refer the reader to the surveys @xcite for a more detailed discussion .
all surfaces considered in this section will be of genus two . given a positive integer @xmath18 let @xmath153 be the order of discriminant @xmath18 in the field @xmath154 .
we say that @xmath41 is an _ eigenform surface of discriminant @xmath18 _ if @xmath153 embeds in the endomorphism ring of its jacobian .
[ thm : mcmullen dynamics h(2 ) ] any @xmath13-orbit in @xmath8 is either closed or dense .
[ thm : calta mcmullen h(2 ) ] the following are equivalent for @xmath155 : * @xmath41 is a lattice surface . *
@xmath41 is an eigenform surface . *
any saddle connection direction is completely periodic . *
any cylinder direction is completely periodic .
suppose the above hold and let @xmath18 be the corresponding discriminant .
then : * @xmath41 is arithmetic if and only if @xmath18 is a square . *
the number of @xmath13-orbits of lattice surfaces of discriminant @xmath15 is : @xmath156 we will refer to the @xmath13-orbits of lattice surfaces in @xmath8 by their discriminant @xmath18 ; in case @xmath157 we will write @xmath158 to distinguish the two cases in ( these are distinguished by the parity of their spin structure , see @xcite ) .
we will abuse notation by referring to the symbols @xmath158 as discriminants .
( 0,0 ) node [ left ] @xmath39 ( 0,1 ) ; ( 0,0 ) node [ left ] @xmath159 ( 0,-1.3028 ) ; ( 1,1 ) node [ above ] @xmath113 ( 4,1 ) ; ( 0,1 ) node [ above ] @xmath112 ( 1,1 ) ; ( 0,-1.3028 ) node [ below ] @xmath159 ( 1.3028,-1.3028 ) ; ( 1,1 ) node [ right ] @xmath148 ( 1,0.3 ) ; ( 2.3028,1 ) node [ right ] @xmath150 ( 2.3028,0.3 ) ; ( 0,0 ) node
node @xmath128 ( 0,-1.3028 ) node
node @xmath127 ( 1.3028,-1.3028 ) node
node @xmath128 ( 1.3028,0 ) node node @xmath126 ( 3,0 )
node @xmath133 ( 4,1 ) node node @xmath126 ( 2.3028,1 )
node @xmath127 ( 1,1 ) node node @xmath133 ( 0,0 ) cycle ; [ fig : prototype ] [ thm : mcmullen prototypes h(2 ) ] suppose @xmath44 is a completely periodic direction on a lattice surface in @xmath8 of discriminant @xmath18 , such that the corresponding cylinder decomposition has two cylinders .
then one of the cylinders @xmath43 is simple , and there are integers @xmath160 such that : @xmath161 after applying an affine map which makes the core curve of @xmath43 horizontal , a saddle connection crossing @xmath43 vertical , and the height of @xmath43 equal to its circumference , and equal to @xmath162 the surface is presented as in figure [ fig : prototype ] .
furthermore , for each @xmath18 there are only finitely many choices of @xmath163 satisfying , and for each such choice , the surface of figure [ fig : prototype ] is a surface of discriminant @xmath18 with a completely periodic horizontal direction . in case @xmath164 ,
the parity of its spin structure is @xmath165 where @xmath166 depends on @xmath18 .
our notation differs slightly from that of mcmullen @xcite , who took @xmath167 and introduced an additional parameter @xmath168 .
our @xmath112 corresponds to @xmath169 in mcmullen s notation .
mcmullen s notation was better suited to his purpose of listing the different prototypes which correspond to the same cusp of the corresponding @xmath13-orbit information which will not be relevant to our discussion .
let @xmath20 be the _ eigenform locus of discriminant @xmath18 _ , i.e. the subset of @xmath1 consisting of eigenform surfaces of discriminant @xmath18 .
we say that two surfaces @xmath170 are _ rel equivalent _ if there is an identification of their homology groups @xmath171 such that corresponding 1-cycles have the same holonomy .
this equivalence relation is generated by a local surgery which moves one singularity with respect to the other without changing absolute periods .
[ sec : unstable convexity ] for more details .
[ thm : calta mcmullen h(1,1 ) ] the eigenform locus @xmath20 is closed and invariant under both the @xmath13-action and rel - equivalence .
the following are equivalent for @xmath135 : * @xmath41 is an eigenform surface . *
any cylinder direction is completely periodic .
[ thm : mcmullen eigenform ] for @xmath172 , each @xmath20 is a connected 5-dimensional suborbifold of @xmath1 which is equipped with a unique smooth @xmath13-invariant ergodic measure .
this measure is finite .
[ thm : mcmullen dynamics h(1,1 ) ] for any surface in @xmath135 , one of the following ( mutually exclusive ) assertions holds : * @xmath41 is an eigenform surface of discriminant @xmath18 and is a lattice surface . in this case
@xmath41 is either arithmetic , and @xmath18 is a square , or @xmath41 is in the orbit of the regular decagon , and @xmath173 .
* @xmath41 is an eigenform surface of discriminant @xmath18 and its @xmath13-orbit is dense in @xmath20 .
* @xmath41 is not an eigenform surface and its @xmath13-orbit is dense in @xmath1 .
given @xmath174 and @xmath175 we write @xmath176 that is @xmath177 is the unique element in @xmath178 the following result is proved in @xcite [ thm : prototypes h(1,1 ) ] suppose @xmath44 is a completely periodic direction on an eigenform surface of discriminant @xmath18 in @xmath1 , such that the corresponding cylinder decomposition has three cylinders , or two cylinders with one simple cylinders .
then there are integers @xmath160 , and real numbers @xmath179 such that @xmath163 satisfy , and @xmath180 after applying an affine map as in theorem [ thm : mcmullen prototypes h(2 ) ] , and setting @xmath181 the surface is presented as in figure [ fig : prototype h(1,1 ) ] .
the number of cylinders in the cylinder decomposition is two if and only if @xmath182 .
furthermore , suppose @xmath183 satisfy and , and define @xmath184 via .
then the surface of figure [ fig : prototype ] is a surface of discriminant @xmath18 with a completely periodic horizontal direction .
\(a ) at ( 0 , 0 ) ; ( h ) at ( 0.2 , 0.6 ) ; ( g ) at ( 1.2 , 0.6 ) ; ( f ) at ( 2.2 , 0.6 ) ; ( e ) at ( 2 , 0 ) ; ( b ) at ( 0 , -1 ) ; ( c ) at ( 1 , -1 ) ; ( d ) at ( 1 , 0 ) ; ( i ) at ( 1.6 , -0.4 ) ; ( j ) at ( 2.6 , -0.4 ) ; \(a ) ( b ) ( c ) ( d ) ( i ) ( j ) ( e ) ( f ) ( h ) ( a ) cycle ; ( g ) node [ right ] @xmath150 + ( 0,-0.2 ) ; ( e ) node [ left ] @xmath151 + ( 0,-0.2 ) ; ( h ) node [ right ] @xmath148 + ( 0,-0.2 ) ; ( 0,0 ) node [ left ] @xmath39 ( 0,0.65 ) ; ( 0,0 ) node [ left ] @xmath159 ( 0,-1 ) ; ( 0.2,0.6 ) node [ above ] @xmath113 ( 2.25,0.6 ) ; ( -0.1,0.6 ) node [ above ] @xmath112 ( 0.2,0.6 ) ; ( 0,-1 ) node [ below ] @xmath159 ( 1,-1 ) ; ( 1,-0.45 ) node [ below ] @xmath185 ( 1.55,-0.45 ) ; ( 1.1,0 ) node [ left ] @xmath168 ( 1.1,-0.4 ) ; ( a ) node ( h ) node ( g ) node
( f ) node ( e ) node
( a ) ( b ) node ( c ) node ( d ) node ( i ) node ( j ) node
( e ) ; [ fig : prototype h(1,1 ) ] note that when @xmath186 there are two simple cylinders in the presentation , the taller one is represented by a @xmath187 square . by an abuse of notation which should cause no confusion
, we will refer to the latter simple cylinder as _ the _ simple cylinder of the prototype .
we say that @xmath41 has a _ one - cylinder presentation _ if it can be presented as one cylinder with gluings along its boundary edges .
the following useful result is well - known to those who know it well . for an elegant proof of hubert ,
see ( * ? ? ?
* lemma 5.3 ) . for a generalization ,
see @xcite .
if @xmath41 is a lattice surface with a one - cylinder presentation then @xmath41 is arithmetic .
we will also need the following statement , which follows from @xcite : [ hubert lelievre mcmullen ] any arithmetic surface in @xmath8 has a one - cylinder presentation .
the conclusion of proposition [ hubert lelievre mcmullen ] fails in @xmath1 . for an example , see the surface portrayed on the left hand side of figure [ fig : lots of squares ] . in recent breakthrough work
@xcite , severe restrictions were found on the possible @xmath13-invariant measures and @xmath13-orbit - closures in any stratum @xmath96 .
we will need the following : [ thm : finiteness emm ] for any stratum @xmath96 , any closed @xmath13-invariant subset is a union of finitely many orbit - closures .
as an immediate consequence of proposition [ prop : g invariant closed ] and theorems [ thm : mcmullen dynamics h(2 ) ] , [ thm : calta mcmullen h(2 ) ] , [ thm : mcmullen dynamics h(1,1 ) ] and [ thm : finiteness emm ] , we have : [ cor : emm finitely many ] in @xmath8 , the set of surfaces with no strictly convex presentations consists of finitely many @xmath13-orbits of lattice surfaces . in @xmath1 , the set of surfaces with no strictly convex presentations consists of finitely many @xmath13-orbits of lattice surfaces and finitely many eigenform loci
. one could derive corollary [ cor : emm finitely many ] without appealing to @xcite , using mcmullen s results @xcite and an analysis of horocycle trajectories close to the supports of the @xmath13-invariant measures of @xmath1 .
we omit the details .
[ sec : nec and suff ] in this section we will start with a surface of genus two presented via the prototypes of [ sec : genus 2 ] , and give an alternative presentation of the same surface via what we will call the _
canonical polygon_. this will be an octagon ( resp .
a decagon ) with edge identifications giving rise to a surface in @xmath8 ( resp .
@xmath1 ) .
the construction of the canonical polygon is explicit and algorithmic .
we will show that if a surface of genus two admits a strictly convex presentation , then it is a canonical polygon for some prototype .
we will discuss the two strata @xmath8 and @xmath1 separately .
[ subsec : canonical h(2 ) ] let @xmath41 be a lattice surface in @xmath8 presented via a prototype as in theorem [ thm : mcmullen prototypes h(2 ) ] , and let @xmath43 be the simple cylinder of the prototype . recall that @xmath43 is normalized so that its circumference and height are both equal to @xmath159 , its circumference is horizontal , and there is a vertical saddle connection passing through @xmath43 from bottom to top .
we will say that the prototype is _
nondegenerate for @xmath41 _ if every non - critical vertical leaf passes through @xmath43 infinitely many times , and every critical leaf passes through the interior of @xmath43 at least once .
otherwise we will say the prototype is _ degenerate for @xmath41_. let @xmath188 be the horizontal saddle connection which is the top boundary segment of @xmath43 .
[ prop : iet ] the prototype with parameters @xmath189 is nondegenerate for @xmath41 if and only if there are non - negative integers @xmath190 and @xmath191 such that the following hold : @xmath192 \ \mathrm { for } \
j=1 , \ldots , k-1,}\ ] ] and @xmath193\ \mathrm{for } \
j=1 , \ldots , \ell-1.}\ ] ] in this case the return map to @xmath188 going upward along vertical leaves is an interval exchange on three intervals , and the corresponding permutation is the ` inversive permutation ' @xmath194 nondegeneracy is equivalent to all minimal components of the vertical flow intersecting the interior of @xmath43 .
equivalently , all vertical separatrices going down from the singularity pass through @xmath43 .
let @xmath190 and @xmath191 denote respectively the number of times the downward vertical separatrices marked @xmath144 in figure [ fig : prototype ] , pass through the non - simple cylinder in the prototype , before reaching the interior of @xmath43 .
the horizontal coordinate of a point changes by an addition of @xmath112 whenever the separatrix passes through the non - simple cylinder without reaching @xmath43 .
this leads to the formulae and .
the two points where the two prongs touch @xmath188 are the two discontinuities of the interval exchange which is the return map to @xmath188 .
thus the interval exchange has two discontinuities , i.e. three intervals of continuity .
the center of @xmath43 is a weierstrass point .
symmetry of the vertical foliation with respect to the hyperelliptic involution implies that the permutation is inversive , as claimed .
the canonical polygon for the prototype is constructed as follows ( see figure [ fig : canonical polygon ] ) : for each of the three intervals of continuity of the interval exchange in proposition [ prop : iet ] , construct the strip going up along vertical leaves from the top of @xmath43 to the bottom of @xmath43 , and number them from left to right .
the singularity of @xmath41 appears on the right - hand boundary edge of the first strip , on both boundary edges of the second strip , and on the left - hand boundary edge of the third strip .
cut the strips by segments joining the singularity to itself , one in each strip , and rearrange .
the canonical polygon is the resulting octagon .
\(a ) at ( 0,0 ) ; ( b ) at ( 0,-1.3028 ) ; ( c ) at ( 1.3028,-1.3028 ) ; ( d ) at ( 1.3028,0 ) ; ( e ) at ( 3,0 ) ; ( f ) at ( 4,1 ) ; ( g ) at ( 2.3028,1 ) ; at ( g ) @xmath195 ; ( h ) at ( 1,1 ) ; ( ge ) at ( 2.1514,0 ) ; ( eg ) at ( 3.1514,1 ) ; ( g1 ) at ( 2.3028,0 . ) ; ( g2 ) at ( 3.3028,1 ) ; ( g3 ) at ( 3.3028,0 ) ; ( g4 ) at ( 0.3028,1 ) ; ( g5 ) at ( 0.3028,0 ) ; ( a ) ( d ) ( h ) cycle ; ( g ) ( ge )
( e ) cycle ; ( e ) ( eg ) ( f ) cycle ; ( h ) ( d ) ; ( d )
( g ) ; ( e ) ( eg ) ; ( ge )
( g ) ; ( g ) ( e ) ; ( a ) ( d ) ; ( h ) ( 1,0 ) ; ( g ) ( g1 ) ; ( g2 ) ( intersection of e f and g2g3 ) ; ( intersection of a h and g4g5 )
( g5 ) ; ( a ) node (b ) node (c ) node (d ) node (e ) node (f ) node (g ) node (h ) node (0,0 ) cycle ; \(a ) at ( 0,0 ) ; at ( a ) @xmath196 ; ( b ) at ( 0,-1.3028 ) ; ( c ) at ( 1.3028,-1.3028 ) ; ( d ) at ( 1.3028,0 ) ; ( g ) at ( 0.3028,2 ) ; at ( g ) @xmath197 ; ( h ) at ( 1,1 ) ; at ( h ) @xmath198 ; ( hh ) at ( 1,1 ) ; ( hh ) at ( 0.3028,-2.3028 ) ; ( gg ) at ( 1,-3.3028 ) ; ( h1 ) at ( 0,1 ) ; ( hh1 ) at ( 1.3028,-2.3028 ) ; ( hl ) at ( intersection of a g and h1h ) ; ( a ) ( d )
( h ) cycle ; ( a )
( hl ) cycle ; ( hl )
( g ) cycle ; ( a )
( h ) ; ( b ) ( c ) ; ( intersection of a g and h1h )
( h ) ; ( intersection of gg c and hh1hh )
( hh ) ; ( c ) ( hh ) ; ( a ) ( d ) ; ( h ) ( 1,0 ) ; ( g ) ( 0.3028,0 . ) ; ( a ) node (b ) node (hh ) node ( gg ) node ( c ) node (d ) node (h ) node (g ) node ( 0,0 ) cycle ; [ fig : canonical polygon ]
let @xmath199 denote respectively the holonomies of the segment going from the top left - hand corner of @xmath43 to the singularity of @xmath41 , by following the top of @xmath43 until the first ( respectively second ) discontinuity of the interval exchange transformation described in proposition [ prop : iet ] , and then up to the singularity along the right hand edge of the first ( resp .
second ) strip .
[ prop : canonical octagon ] let @xmath190 and @xmath191 be as in and
. then @xmath200 and @xmath201 .
the canonical polygon is strictly convex if and only if @xmath202 the points @xmath203 are vertices of the octagon in light of proposition [ prop : iet ] . to find out which of these two points is to the left of the other ,
we check how these vertices are identified with the vertices in the prototype diagram ; we see that in the octagon , the vertical prong going down from the vertex at @xmath204 is reached from the vertical prong at @xmath196 by moving counterclockwise for an angle of @xmath54 .
the same description is valid for the vertical prong at @xmath205 , thus they represent the same prong .
the two inequalities in correspond to internal angles less than @xmath9 at each of the two vertices @xmath199 .
the other six vertices follow automatically from these via the hyperelliptic involution , and from the inequalities @xmath206 .
our interest in the canonical polygon is motivated by the following : [ prop : nec and suff ] if @xmath41 is a lattice surface in @xmath8 and has a strictly convex presentation , then there is a prototype which is nondegenerate for @xmath41 , for which the given convex polygon presentation coincides with the canonical polygon .
let @xmath207 be the convex polygon representing @xmath41 and let @xmath43 be a simple cylinder as in proposition [ prop : central symmetry ] .
then the direction of the circumference of @xmath43 is completely periodic by theorem [ thm : calta mcmullen h(2 ) ] .
consider the prototype corresponding to this direction via theorem [ thm : mcmullen prototypes h(2 ) ] .
the convexity of @xmath207 implies that the prototype is nondegenerate for @xmath41 .
the definition of @xmath208 given in the discussion preceding proposition [ prop : canonical octagon ] now shows that @xmath207 coincides with the canonical polygon for this prototype .
[ subsec : canonical h(1,1 ) ] we proceed analogously to [ subsec : canonical h(2 ) ] , making the required adjustments .
let @xmath41 be an eigenform surface in @xmath8 .
suppose @xmath41 is presented via a prototype as in theorem [ thm : prototypes h(1,1 ) ] , and let @xmath43 be the simple cylinder of the prototype .
as before , we will say that the prototype is _
nondegenerate for @xmath41 _ if every non - critical vertical leaf passes through @xmath43 infinitely many times , and every critical leaf passes through the interior of @xmath43 at least once , and _ degenerate _ otherwise .
note that the vertical direction need no longer be a completely periodic direction for @xmath41 .
let @xmath188 be the horizontal saddle connection which is the top boundary segment of @xmath43 .
let @xmath209 be as in , define @xmath184 via , and define @xmath210 \to [ 0,b ) , \ \
r(z ) { { \ , \stackrel{\mathrm{def}}{=}\ , } } r_b(z - x+s ) . }
\ ] ] [ prop : iet2 ] the prototype with parameters @xmath211 is nondegenerate for @xmath41 if and only if @xmath212 and there are non - negative integers @xmath213 such that the following hold : @xmath214 \ \mathrm { for } \ j=1 , \ldots , k-1,}\ ] ] @xmath215\ \mathrm{for } \ j=1 , \ldots , \ell-1}\ ] ] and @xmath216\ \mathrm{for } \ j=1 , \ldots , m-1 . }
\ ] ] in this case the return map to @xmath188 going upward along vertical leaves is an interval exchange on four intervals , and the corresponding permutation is the ` inversive permutation ' @xmath217 the map @xmath218 in is the return map to the bottom of the non - simple cylinder of the prototype , for those @xmath219 which do not begin along the boundary with the simple cylinder @xmath43 .
as before , nondegeneracy is equivalent to the requirement that all vertical separatrices going down from the singularity pass through the interior of @xmath43 .
let @xmath213 denote respectively the number of times the downward vertical separatrices @xmath220 respectively in figure [ fig : canonical polygon h(1,1 ) ] , pass through the non - simple cylinder in the prototype , before reaching the interior of @xmath43 .
this leads to the formulae , and .
the three points in the interior of @xmath188 are the three discontinuities of the interval exchange which is the return map to @xmath188 .
thus the interval exchange has four intervals of continuity .
counting weierstrass points one sees that there must be a weierstrass point in the center of @xmath43 . in particular @xmath43 is fixed by the involution .
symmetry of the vertical foliation with respect to the hyperelliptic involution implies that the permutation is inversive , as claimed .
\(a ) at ( 0 , 0 ) ; ( h ) at ( 1.4196 , 0.6845 ) ; ( g ) at ( 3.151666 , 0.6845 ) ; ( f ) at ( 4.4196 , 0.6845 ) ; ( e ) at ( 3 , 0 ) ; ( b ) at ( 0 , -1.732 ) ; ( c ) at ( 1.732 , -1.732 ) ; ( d ) at ( 1.732 , 0 ) ; ( i ) at ( 2.63205 , -0.2 ) ; ( j ) at ( 3.9 , -0.2 ) ; \(a ) ( b ) ( c ) ( d ) ( i ) ( j ) ( e ) ( f ) ( h ) ( a ) cycle ; ( a ) ( d ) ( h ) cycle ; ( e ) ( f ) ( g ) cycle ; ( i ) ( j ) ( e ) cycle ; ( h )
( d ) ; ( d ) ( e ) ; ( e ) ( g ) ; ( a ) ( d ) ; ( i ) ( e ) ; ( g ) + ( 0,-0.6845 ) + + ( -3,0 ) + ( 0,-0.6845 ) ; ( h ) + ( 0,-0.6845 ) ; ( e )
+ ( 0,-0.6845 ) + + ( 0.5196,0.6845 )
+ ( 0,-0.6845 ) ( .5196,0.6845 )
+ ( 0,-0.6845 ) ; \(a ) node ( h ) node ( g ) node
( f ) node ( e ) node ( j ) node ( i ) node ( d ) node ( c ) node ( b ) node ( a ) cycle ; \(a ) at ( 0 , 0 ) ; ( h ) at ( 1.4196 , 0.6845 ) ; ( g ) at ( 0.151666 , 0.6845 ) ; ( f ) at ( 4.4196 , 0.6845 ) ; ( e ) at ( 0.5196 , 0.8845 ) ; ( b ) at ( 0 , -1.732 ) ; ( c ) at ( 1.732 , -1.732 ) ; ( d ) at ( 1.732 , 0 ) ; ( i ) at ( 2.63205 , -0.2 ) ; ( j ) at ( 3.9 , -0.2 ) ; ( gg ) at ( 1.580334 , -2.4165 ) ; ( ee ) at ( 1.212334 , -2.6141 ) ; ( hh ) at ( 0.3124 , -2.4165 ) ; \(a ) ( g ) ( e ) ( h )
( d ) ( c ) ( gg ) ( ee ) ( hh ) ( b ) ( a ) cycle ; ( a ) ( d ) ( h ) cycle ; ( a ) ( h )
( g ) cycle ; ( g )
( e ) cycle ; ( g ) ( h ) ; ( a ) ( h ) ; ( a ) ( d ) ; ( b ) ( c ) ; ( gg ) ( hh ) ; ( g )
+ ( 0,-0.6845 ) ; ( h ) + ( 0,-0.6845 ) ; ( e ) + ( 0,-0.8845 ) ; \(a ) node ( g ) node ( e ) node ( h ) node ( d ) node ( c ) node ( gg ) node ( ee ) node ( hh ) node ( b ) node ( a )
cycle ; [ fig : canonical polygon h(1,1 ) ] the canonical polygon for the prototype is constructed as before .
it is a decagon .
let @xmath221 denote respectively the holonomies of the segment going from the top left - hand corner of @xmath43 to the singularity of @xmath41 , by following the top of @xmath43 until the first ( respectively second and third ) discontinuity of the interval exchange transformation described in proposition [ prop : iet ] , and then up to the singularity along the right hand edge of the first ( resp .
second and third ) strip .
the following propositions are proved analgously to propositions [ prop : canonical octagon ] and [ prop : nec and suff ] .
we leave the proofs to the reader . [
prop : canonical decagon ] let @xmath190 and @xmath191 be as in and
. then @xmath222 and @xmath223 .
the canonical polygon is strictly convex if and only if @xmath224 [ prop : nec and suff2 ] if @xmath41 is a lattice surface in @xmath1 and has a strictly convex presentation , then there is a prototype which is nondegenerate for @xmath41 , for which the given convex polygon presentation coincides with the canonical polygon .
the main result of this section is : [ prop : for main ] the list is the complete list of discriminants for which there is no prototype for which the canonical polygon is strictly convex .
the stratum @xmath8 contains a surface with a strictly convex presentation , viz . the regular octagon .
therefore there is an open subset of @xmath225 of @xmath8 consisting of surfaces with a strictly convex presentation . by theorem [ thm : mcmullen dynamics h(2 ) ] , if @xmath41 is not a lattice surface then its @xmath13-orbit intersects @xmath225 and therefore @xmath41 has a strictly convex presentation
. therefore the first assertion of theorem [ thm : main ] follows from propositions [ prop : nec and suff ] and [ prop : for main ] . for the second assertion , note that is obtained from by removing all @xmath18 which are square .
thus it suffices to show that if @xmath18 is a square , @xmath41 has a convex presentation , and if @xmath18 is not a square and the surface of discriminant @xmath18 has a convex presentation , it must be strictly convex . the first part of this claim follows from proposition [ hubert lelievre mcmullen ] . now suppose @xmath41 has a convex presentation which is not strictly convex , then the corresponding polygon @xmath207 must be either a parallelogram or a hexagon with some vertices in the interiors of edges .
if @xmath207 is a parallelogram then the surface has a one - cylinder presentation , and @xmath18 is a square by proposition [ hubert lelievre mcmullen ] . if @xmath207 is a hexagon , let @xmath44 be the direction of a side of @xmath207 .
the flow in direction @xmath44 is completely periodic by theorem [ thm : calta mcmullen h(2 ) ] , and the return map to a diagonal of the hexagon is an interval exchange on two intervals .
this means that this return map is a rational rotation , i.e. @xmath207 maps to a one - cylinder presentation with waist curve in direction @xmath44 . again using proposition [ hubert lelievre mcmullen ]
we find that @xmath18 is a square .
given even @xmath18 let @xmath226 be such that @xmath227 set @xmath228 and @xmath229 so that @xmath230 now define @xmath231 so that @xmath232 let @xmath112 be the unique integer in the interval @xmath233 [ prop : computations , even ] the following hold when @xmath234 : * @xmath235 * @xmath236 .
* @xmath237 .
* @xmath238 . *
@xmath239 * @xmath240 * @xmath241 it follows from and that @xmath242 which implies ( i ) , since @xmath243 to prove ( ii ) , define @xmath244 we need to show @xmath245 plugging in @xmath246 and @xmath247 , this is equivalent to @xmath248,}\ ] ] which is easily proved using the quadratic formula .
the right hand side of ( iii ) follows from ( ii ) and the definition of @xmath112 .
the left hand side follows from the definition of @xmath112 and ( i ) . since @xmath249 , ( iv ) follows from the definition of @xmath112 .
the inequality ( v ) follows from ( ii ) and ( iv ) , and inequality ( vi ) follows from ( iii ) and ( v ) . inequality ( vii ) is immediate from the definition of @xmath112 . given odd @xmath18 let @xmath226 be such that @xmath250 set @xmath251 and define @xmath252 via , using @xmath253 respectively , so that @xmath254 define @xmath255 via so that @xmath256 let @xmath257 be the unique integer in the interval @xmath258 and let @xmath259 be the unique integer in the interval @xmath260 . [
prop : computations , odd ] inequalities ( i)(vii ) from proposition [ prop : computations , even ] hold for @xmath261 when @xmath234 . if @xmath262 then : * @xmath263 * @xmath264 . * @xmath265 . * @xmath266 . * @xmath267 * @xmath268 * @xmath269 very similar to the proof of proposition [ prop : computations , even ] , and we leave the details to the reader .
[ cor : d big ] for any @xmath270 the lattice surface of discriminant @xmath18 has a strictly convex presentation .
suppose first that @xmath18 is even .
let @xmath271 be as in proposition [ prop : computations , even ] , and set @xmath272 then one easily checks that holds so that @xmath273 is a prototype
. inequalities ( v ) , ( iii ) , ( iv ) and ( vii ) of proposition [ prop : computations , even ] show that @xmath274 satisfies , and inequalities ( iii ) and ( vi ) imply that @xmath275 satisfies .
finally , follows from ( ii ) .
if @xmath18 is odd we can use either of the prototypes @xmath276 or @xmath277 , where in the first case we use @xmath278 and in the second we use @xmath279 it is easy to check using proposition [ prop : computations , odd ] that the hypotheses of proposition [ prop : canonical octagon ] are satisfied . if @xmath280 then this concludes the proof .
if @xmath281 there are two possible spins , so it remains to check that the two cases @xmath282 and @xmath277 correspond to different spins , and this is immediate from @xmath283 and .
[ subsec : computer algorithm ] in order to deal with the remaining cases @xmath284 , we have implemented the following algorithm : 1 .
given @xmath18 , enumerate all solutions @xmath163 to .
2 . for each prototype @xmath163 from step 1 , do : 1 .
let @xmath190 be the smallest @xmath68 so that @xmath285 $ ] , and let @xmath191 be the smallest @xmath68 so that @xmath286 $ ] ( cf . and ) .
2 . if @xmath287 then the prototype is degenerate for @xmath41 .
3 . check whether holds .
if yes , stop : the canonical octagon for this prototype is a strictly convex presentation .
if fails for all prototypes , then there is no strictly convex presentation .
we have implemented this algorithm on sage .
the code and output of our computation are available along with this paper at @xcite .
a sample of the output of our program is below in figure [ fig : sample output ] .
using the computation we find : [ prop : computer ] for @xmath284 , there is a prototype with discriminant @xmath18 whose canonical octagon is convex , if and only if @xmath18 does not appear on the list .
immediate from corollary [ cor : d big ] and proposition [ prop : computer ] . , both spins .
the boxed @xmath288 s are the convex presentations of corollary [ cor : d big ] ] [ fig : sample output ] assertions ( i ) and ( ii ) were proved in
[ sec : 9 and 16 ] and [ sec : finding ] respectively . to deduce ( iii ) from ( ii ) , note that if @xmath18 is not a square , and is not equal to 5 , by theorem [ thm : mcmullen dynamics h(1,1 ) ] , any @xmath13-orbit in @xmath20 is dense in @xmath20 . moreover in case @xmath173 the only orbit which is not dense in @xmath437 is that of the regular decagon , which is clearly strictly convex .
thus the @xmath13-orbit of any surface which is neither arithmetic , nor contained in @xmath417 , contains a surface with a strictly convex presentation .
this concludes the proof in light of @xmath13-invariance .
assertion ( iv ) follows from corollary [ cor : emm finitely many ] . to prove ( v ) note that the collections of lattice surfaces in @xmath417 consists of a countable collection of @xmath13-orbits , so in particular is of measure zero . in view of corollary
[ cor : measure zero ] , almost every surface in @xmath417 has the required properties .
[ sec : 9 and 16 ] [ prop : 4 9 16 ] no surface in @xmath24 has a strictly convex presentation . surfaces in @xmath20 with @xmath290 admit a degree @xmath291 cover onto a torus with two branch points , corresponding to the distinct singularities of the surface .
suppose @xmath292 has a convex presentation and , applying the @xmath13-action and dilation , suppose it covers the standard square once - punctured torus with holonomy in @xmath293 , via a degree @xmath291 covering map .
let @xmath49 be the convex decagon representing @xmath41 , label its vertices cyclically counterclockwise , and let @xmath294 be the pentagon which is the convex hull of vertices labeled 1,3,5,7,9 .
then the vertices of @xmath294 all project to the same point on the torus .
this implies that @xmath294 is a strictly convex pentagon in the plane , with vertices in @xmath293 and area strictly less than @xmath291 .
a further restriction on @xmath294 is that the sum of the internal angles at adjacent vertices of @xmath294 ( i.e. vertices whose labeling on @xmath49 differs by 2 ) is strictly greater than @xmath9 .
this follows from an argument as in the last paragraph of the proof of proposition [ prop : central symmetry ] .
we call this the _ adjacent angles restriction . _ thus to conclude the proof it suffices to show that for @xmath295 , there is no @xmath294 with the stated properties .
note that all properties are invariant under the action of @xmath296 integer matrices of deteminant @xmath297 .
suppose first that there are 2 adjacent edges of @xmath294 which do not contain lattice points on their interior . by applying an integer matrix
we can assume that three of the vertices of @xmath294 are @xmath298 .
we write the two additional vertices @xmath299 as @xmath300 and @xmath301 respectively . then applying the adjacent angles restrictions on the various edges obtain the following inequalities : @xmath302 from these we derive @xmath303 which is impossible . for the remaining case we will use pick s formula again ( see the proof of proposition [ prop : interesting square tiled ] )
. if no two adjacent edges of @xmath294 are without interior lattice points , then the number of lattice points is at least 8 ( 5 at the vertices and at least 3 on the edges ) .
additionally , we can apply an integer matrix to assuming that three of the adjacent vertices of @xmath294 are @xmath304 with @xmath305 and @xmath306 .
then the adjacent angle restrictions imply that the points @xmath152 and @xmath307 are interior points of @xmath294 .
thus by pick s formula the area of @xmath294 is at least 5 .
the case @xmath308 could also be proved by using corollary [ cor : short saddle connection ] and noting that all surfaces in @xmath2 have nontrivial translation automorphisms .
it is possible to extend proposition [ prop : 4 9 16 ] to higher genus , as follows . for @xmath29 , and @xmath309 , let @xmath310 denote the hurwitz space of surfaces in the hyperelliptic component of the stratum @xmath11 which admit a degree @xmath291 cover of a torus branched over two points .
proposition [ prop : 4 9 16 ] corresponds to the case @xmath311 of the following statement .
[ remark : more general ] for any @xmath312 there is @xmath313 such that for @xmath314 , @xmath310 contains no surface with a strictly convex presentation .
we first note that the restriction @xmath315 is imposed in order to ensure that @xmath316 indeed , the hurwitz formula implies that @xmath317 when @xmath318 , and it is not hard to see using ` staircases ' that the loci @xmath310 are nonempty for @xmath315 .
we will prove the statement for @xmath319 a convex presentation , if it exists in @xmath310 , is a @xmath320-gon .
arguing as in the proof of proposition [ prop : 4 9 16 ] , we find that it contains a strictly convex @xmath321-gon @xmath294 with vertices at lattice points , and with area strictly less than @xmath291 .
moreover @xmath294 satisfies the following restriction ( generalizing the adjacent edges restriction ) : let @xmath322 denote the holonomies of the edges of @xmath294 , ordered cyclically counterclockwise and equipped with the boundary orientation of @xmath294 . as usual
we consider indices mod @xmath323 .
then for @xmath324 and for any @xmath325 , the angle between @xmath326 and @xmath327 ( measured counterclockwise ) is less than @xmath9 , and if @xmath323 is even , the angle between @xmath326 and @xmath328 is @xmath9 .
we need to show that no such @xmath294 exists .
let @xmath329 denote the minimal area of a strictly convex @xmath321-gon with vertices in @xmath293 .
then it is known ( see @xcite ) that @xmath330 and @xmath331 for @xmath332 .
so there is nothing to prove for @xmath332 . if @xmath333 we need to show any strictly convex @xmath321-gon with vertices in @xmath293 , of area less than @xmath334 , does not satisfy the generalized adjacent edges restriction .
this can be done by a case - by - case analysis involving pick s formula , as in the proof of proposition [ prop : interesting square tiled ] ; we leave the details to our dedicated reader .
[ sec : finding ] one can consider @xmath8 as part of a bordification of @xmath1 , where the boundary added corresponds to surfaces obtained by ` collapsing the two singularities . ' since having a strictly convex presentation is an open condition , one may expect that having a strictly convex presentation in @xmath8 implies having one nearby , in @xmath1 .
indeed this turns out to be the case , as we will show in this section . for a discussion of this point of view , and a careful discussion of the inverse operation of ` splitting apart a zero ' , see @xcite .
let @xmath49 be a strictly convex centrally symmetric octagon with vertices @xmath335 , listed counterclockwise , and let @xmath41 denote the surface in @xmath8 obtained by gluing opposite sides of @xmath49
. denote by @xmath336 the holonomies of saddle connections representing edges of @xmath49 ( where indices are considered mod @xmath337 ) .
let @xmath338 be a vector satisfying @xmath339 where @xmath340 denotes the determinant of the matrix with column vectors @xmath341 ; in other words , @xmath338 lies in the convex cone generated by @xmath342 and @xmath343 . for any @xmath344 , let @xmath345 denote the centrally symmetric decagon whose vertices ( listed counterclockwise ) are the points @xmath346 that is we obtained @xmath347 by splitting each of the vertices @xmath348 in two and replacing them by two nearby vertices separated by motion in the direction of @xmath338 ; and all other vertices have been moved at most a distance @xmath349 from their initial positions .
since @xmath49 is strictly convex , the choice of direction of @xmath338 ensures that there exists @xmath350 such that for any @xmath351 , @xmath347 is also strictly convex .
see figure [ fig : octagon - decagon ] .
( 0,-2.50 ) node [ inner sep=0pt ] ( v0 ) + ( 0.5,0 ) node [ inner sep=0pt ] ( w0 ) ( 2,-1.725 ) node [ inner sep=0pt ] ( v1 ) ( 2.44,-0.2 ) node [ inner sep=0pt ] ( v2 ) + ( 0.5,0 ) node [ inner sep=0pt ] ( w2 ) ( 1.76,1.755 ) node [ inner sep=0pt ] ( v3 ) ( 0,2.5 ) node [ inner sep=0pt ] ( v4 ) + ( 0.5,0 ) node [ inner sep=0pt ] ( w4 ) ( -2,1.725 ) node [ inner sep=0pt ] ( v5 ) + ( 0.5,0 ) node [ inner sep=0pt ] ( w5 ) ( -2.44,0.2 ) node [ inner sep=0pt ] ( v6 ) ( -1.76,-1.755 ) node [ inner sep=0pt ] ( v7 ) + ( 0.5,0 ) node [ inner sep=0pt ] ( w7 ) ; ( v0 ) node [ below=3pt ] @xmath352 ( v1 ) ( v2 ) ( v3 )
( v4 ) node [ above=3pt ] @xmath353 ( v5 ) ( v6 ) ( v7 ) ( v0 ) cycle ; ( v0 ) + + ( -2 * 0.6,-0.775 * 0.6 ) ; ( v0 ) + + ( 1.76 * 0.6,-0.745 * 0.6 ) ; ( v4 ) + + ( 2 * 0.6,0.775 * 0.6 ) ; ( v4 ) + + ( -1.76 * 0.6,0.745 * 0.6 ) ; ( v0 ) ( w0 ) ( v2 ) ( w2 ) ( v4 ) ( w4 ) ( v5 ) ( w5 ) ( v7 ) ( w7 ) ;
( v0 ) node [ circle , draw , thick , fill = white , inner sep=3pt ] ( v0 ) node [ dart , draw , thick , fill = white , inner sep=1pt ] ( w0 ) node [ dart , draw , thick , fill = white , inner sep=1pt ] ( v1 ) node [ circle , draw , thick , fill = white , inner sep=3pt ] ( v2 ) node [ dart , draw , thick , fill = white , inner sep=1pt ] ( w2 ) node [ dart , draw , thick , fill = white , inner sep=1pt ] ( v3 ) node [ circle , draw , thick , fill = white , inner sep=3pt ] ( v4 ) node [ circle , draw , thick , fill = white , inner sep=3pt ] ( v4 ) node [ dart , draw , thick , fill = white , inner sep=1pt ] ( w4 ) node [ dart , draw , thick , fill = white , inner sep=1pt ] ( v5 ) node [ dart , draw , thick , fill = white , inner sep=1pt ] ( w5 ) node [ dart , draw , thick , fill = white , inner sep=1pt ] ( v6 ) node [ circle , draw , thick , fill = white , inner sep=3pt ] ( v7 ) node [ dart , draw , thick , fill = white , inner sep=1pt ] ( w7 ) node [ dart , draw , thick , fill = white , inner sep=1pt ] ; ( 0,-2.50 ) node [ inner sep=0pt ] ( v0 ) + ( 0.5,0 ) node [ inner sep=0pt ] ( w0 ) ( 2,-1.725 ) node [ inner sep=0pt ] ( v1 ) ( 2.44,-0.2 ) node [ inner sep=0pt ] ( v2 ) + ( 0.5,0 ) node [ inner sep=0pt ] ( w2 ) ( 1.76,1.755 ) node [ inner sep=0pt ] ( v3 ) ( 0,2.5 ) node [ inner sep=0pt ] ( v4 ) + ( 0.5,0 ) node [ inner sep=0pt ] ( w4 ) ( -2,1.725 ) node [ inner sep=0pt ] ( v5 ) + ( 0.5,0 ) node [ inner sep=0pt ] ( w5 ) ( -2.44,0.2 ) node [ inner sep=0pt ] ( v6 ) ( -1.76,-1.755 ) node [ inner sep=0pt ] ( v7 ) + ( 0.5,0 ) node [ inner sep=0pt ] ( w7 ) ; ( v0 ) node [ below=3pt ] @xmath352 ( w0 ) node [ below=3pt ] @xmath354 ( v1 )
( w2 ) ( v3 )
( w4 ) node [ above=3pt ] @xmath355 ( v4 ) node [ above=3pt ] @xmath353
( v6 ) ( w7 ) ( v0 ) cycle ; ( v0 ) ( v1 ) ( v2 )
( v3 ) ( v4 ) ( v5 )
( v6 ) ( v7 ) ( v0 ) cycle ; ( v0 ) ( w0 ) ( v2 ) ( w2 ) ( v4 ) ( w4 ) ( v5 ) ( w5 ) ( v7 ) ( w7 ) ; ( v0 ) node [ circle , draw , thick , fill = white , inner sep=3pt ] ( w0 ) node [ dart , draw , thick , fill = white , inner sep=1pt ] ( v1 ) node [ circle , draw , thick , fill = white , inner sep=3pt ] ( v2 ) node [ dart , draw , thick , fill = white , inner sep=1pt ] ( w2 ) node [ dart , draw , thick , fill = white , inner sep=1pt ] ( v3 ) node [ circle , draw , thick , fill = white , inner sep=3pt ] ( v4 ) node [ circle , draw , thick , fill = white , inner sep=3pt ] ( w4 ) node [ dart , draw , thick , fill = white , inner sep=1pt ] ( v5 ) node [ dart , draw , thick , fill = white , inner sep=1pt ] ( w5 ) node [ dart , draw , thick , fill = white , inner sep=1pt ] ( v6 ) node [ circle , draw , thick , fill = white , inner sep=3pt ] ( v7 ) node [ dart , draw , thick , fill = white , inner sep=1pt ] ( w7 ) node [ dart , draw , thick , fill = white , inner sep=1pt ] ; now let @xmath356 be the surface obtained by identifying opposite sides of @xmath347 .
according to ( * ? ? ?
5.6 ) , the question of whether a genus two surface is an eigenform of discriminant @xmath18 depends only on the absolute period map @xmath357 obtained by integrating the planar 1-forms @xmath358 on @xmath41 .
note that for any @xmath344 , the absolute period map of @xmath359 is the same as that of @xmath41 .
therefore @xmath360 is an eigenform of discriminant @xmath18 whenever @xmath41 is .
this implies : [ prop : oct to dec ] for any @xmath18 , if there is a lattice surface in @xmath8 with discriminant @xmath361 or @xmath362 , with a strictly convex presentation , then there is an eigenform surface of discriminant @xmath18 in @xmath1 with a strictly convex presentation .
[ prop : remaining cases ] for the following values of @xmath18 , the eigenform locus @xmath363 contains surfaces with a strictly convex presentation : @xmath364 note that there are two differences between and : discriminants 9 and 16 appear in and not in ; and all cases of the form @xmath158 are written as @xmath18 . for the case @xmath173 we have the regular decagon .
next we deal with the cases @xmath365 . in these cases , one of the symbols @xmath158 appears in the list but not the other .
for example , contains @xmath366 but not @xmath367 .
suppose with no loss of generality that @xmath368 is not on the list , i.e. there is a lattice surface in @xmath8 corresponding to discriminant @xmath368 , with a strictly convex presentation .
applying proposition [ prop : oct to dec ] we conclude that @xmath20 contains a surface with a strictly convex presentation . for the nine remaining cases
we have found strictly convex presentations by a computer search .
namely we have written a computer program , similar to the one described in [ subsec : computer algorithm ] , which , given a prototype @xmath183 in @xmath1 , checks whether the canonical decagon for this prototype is convex .
this algorithm is based on the discussion of [ subsec : canonical h(1,1 ) ] . for each fixed @xmath18 , there are finitely many possibilities for the discrete values @xmath163 of the prototype . for each fixed value of these , the computer program searched with a large finite number of evenly spaced values of @xmath146 and @xmath145 and stopped when a strictly convex canonical decagon was found . parameters for strictly convex polygons obtained by this computer search are shown in figure [ table h(1,1 ) ] .
the computer code and results of computations can be accessed at @xcite . [ cols= "
> , > , > , > , > , > , > " , ] [ table h(1,1 ) ]
[ sec : unstable convexity ] we have observed that the set of surfaces admitting a strictly convex presentation is open . in this section
we will prove a partial converse for the eigenform loci @xmath369 and @xmath4 .
we will show that if @xmath41 is represented by a decagon @xmath49 which is convex but not strictly convex , then for almost every surface @xmath370 which is sufficiently close to @xmath41 , the corresponding polygon @xmath138 is nonconvex .
we begin by making this precise .
recall that for any @xmath18 , the eigenform locus @xmath20 is a closed suborbifold of @xmath1 , of real dimension 5 .
a _ marked decagon _ is a decagon ( not necessarily convex ) equipped with a partition of its edges into pairs , such that paired edges are parallel and of equal length .
there is an obvious topology on the space of marked decagons . identifying paired edges maps each marked decagon @xmath49 to a translation surface @xmath371 , and
this map @xmath372 is a homeomorphism between a neighborhood of @xmath41 in @xmath1 and a neighborhood of @xmath49 in the space of marked decagons .
we will denote this neighborhood by @xmath373 , and denote the inverse map by @xmath374 .
thus , if @xmath375 and @xmath376 is close to @xmath41 , then @xmath377 is close to @xmath49 and @xmath378 .
[ prop : unstable decs ] suppose @xmath379 or @xmath380 .
suppose @xmath49 is a convex marked decagon such that @xmath381 and @xmath373 is as above .
then there is a neighborhood @xmath382 of @xmath41 such that @xmath383 is a four - dimensional submanifold .
we preface the proof of this proposition with some remarks about the structure of strata and the structure of the loci @xmath20 .
more details can be found in @xcite and @xcite .
let @xmath384 be a model compact orientable surface of genus two with a set @xmath385 consisting of two distinguished points of @xmath384 .
the stratum @xmath1 is locally modelled on the cohomology group @xmath386 , and the eigenform locus @xmath20 is modelled on the semi - direct product group @xmath387 ( where @xmath388 represents motion in the rel leaves ) .
more precisely , for any @xmath135 there is a neighborhood @xmath225 of @xmath41 in @xmath1 , and a homeomorphism of @xmath41 and @xmath384 mapping the singularities of @xmath41 to @xmath389 , such that each @xmath376 is uniquely determined by the holonomies @xmath390 obtained by integrating the translation structure along @xmath391 , for any curve @xmath391 in @xmath384 representing a relative one - cycle in @xmath392 .
moreover ( making @xmath225 smaller if necessary ) , there is a small neighborhood @xmath393 of the identity in @xmath387 , such that for any @xmath394 there are unique @xmath395 , such that for any 1-cycle @xmath391 , @xmath396 here @xmath397 denotes the image of @xmath398 under the linear map @xmath334 .
we will denote the surface corresponding to @xmath399 as above by @xmath400 .
the idea of proof will be to show that given any @xmath49 , there is a one - dimensional linear subspace @xmath401 , such that if @xmath402 is as above , and @xmath403 , then the decagon @xmath404 corresponding to @xmath405 is not convex .
that is , all motions in the rel leaf except possibly motions in a particular direction destroy convexity .
we proceed to the details .
we take @xmath406 small enough so that it is contained in the neighborhood as in the preceding paragraph .
that is , the notation @xmath407 makes sense in @xmath406 , for @xmath408 , and so does the notation @xmath404 .
it follows from the preceding discussion that @xmath409 is obtained from @xmath410 as follows .
we first subdivide the vertices of @xmath49 into two classes , according to whether they map to the singularity labelled @xmath411 or @xmath412 in @xmath41 .
we then keep the vertices mapping to @xmath411 fixed and move all of the others by adding @xmath413 . connecting the vertices
, this defines a new polygon @xmath414 if @xmath413 is sufficiently small ( and we can assume this by making @xmath406 smaller ) .
we then apply the map @xmath334 to @xmath414 .
the resulting polygon is @xmath415 . since linear maps in @xmath13 preserve convexity , in order to prove the proposition it suffices to show that for any convex marked decagon @xmath49 , for which @xmath416 , the set of @xmath413 for which @xmath414 is convex , are contained in a one - dimensional linear subspace . to show this
, it suffices to show that for any @xmath49 , there are three or more edges of @xmath49 which are consecutive and parallel , and connect distinct singularities ; for then any rel perturbation @xmath413 which is not in the direction of these edges will result in @xmath414 being non - convex .
e.g. in the last polygon in figure [ marked decagons ] ( a square ) , this occurs in the top and bottom of the figure .
if we move the black points up ( respectively down ) with respect to the white points , convexity will be destroyed at the vertex between edges @xmath113 and @xmath39 ( respectively between @xmath112 and @xmath113 ) in the top of the figure , so any rel operation which does not move points horizontally will destroy convexity . .
the top row can not appear in @xmath417 . ]
the proof of existence of three parallel consecutive edges proceeds by a case - by - case analysis .
we will show that the only marked decagons which represent surfaces in @xmath1 are those shown in figures [ marked decagons ] and [ marked decagons1 ] .
since , by [ sec : 9 and 16 ] , surfaces in @xmath417 have no strictly convex presentations , the decagon which is the first polygon in figure [ marked decagons ] does not appear in these eigenform loci .
similarly , the octagon and hexagon in the top row of figure [ marked decagons ] also do not appear in @xmath417 , since they can be made into a strictly convex decagon by a rel perturbation , which would again contradict [ sec : 9 and 16 ] .
all the other polygons in the figure satisfy our claim .
this will conclude the proof . .
] when realized as a convex polygon in the plane , some sides of the marked decagon become parallel .
this gives a partition of the set of edges into subsets of parallel edges , where each subset is composed of some pairs of edges identified in the marked decagon ( but possibly more than one pair ) .
considering the different possibilities for these partitions , one finds the following possibilities : * a strictly convex decagon . * a strictly convex octagon in which a pair of parallel edges are each a concatenation of two consecutive edges of @xmath49 .
* a strictly convex hexagon in which two pairs of parallel edges are each a concatenation of two consecutive edges of @xmath49 . * a strictly convex hexagon in which a pair of parallel edges are each a concatenation of three consecutive edges of @xmath49 . * a parallelogram in which a pair of parallel edges is each a concatenation of four consecutive edges of @xmath49 . *
a parallelogram in which one pair of parallel edges is each a concatenation of three consecutive edges of @xmath49 , and one pair of parallel edges is each a concatenation of two consecutive edges of @xmath49
. we will go over these cases one by one and show that all markings which give rise to a surface in @xmath1 appear in figures [ marked decagons ] and [ marked decagons1 ] .
note that the total cone angle around singularities for a surface in @xmath1 is @xmath418 and hence all vertices of the decagon must be singularities of the translation surface .
all indices in the discussion below will be given in cyclic order along the boundary of @xmath49 .
also note that if @xmath419 are consecutive and @xmath252 are consecutive and parallel then we can not glue @xmath257 to @xmath420 and @xmath259 to @xmath421 because then the vertex between @xmath257 and @xmath259 will have cone angle @xmath54 , i.e. will not be a singularity , and we will not be in @xmath1 .
we will call this the _ basic restriction_. clearly case ( a ) corresponds to the first polygon in figure [ marked decagons ] . in case ( b ) , suppose @xmath419 and @xmath252 are two pairs of consecutive edges whose concatenations are parallel edges of the octagon . if @xmath257 is glued to @xmath421 and @xmath259 to @xmath420 in @xmath49 then we have the second polygon in figure [ marked decagons ] .
the other case is forbidden by the basic restriction . by the same reasoning , in case ( c ) , we can only glue sides as in the third polygon in figure [ marked decagons ] . in case ( d ) , let the three pairs of consecutive edges be labelled @xmath422 and @xmath423 .
the gluing in figure [ marked decagons ] corresponds to identifying @xmath424 with @xmath425 .
if @xmath257 is identified with @xmath426 then by the basic restriction we must identify @xmath259 with @xmath420 and @xmath427 with @xmath421 .
now consider the vertex on @xmath427 which is not adjacent to @xmath259 . following the identifications we see that it is identified with three other vertices and has a total angle of @xmath9 , so we are not in @xmath1 . by similar reasoning @xmath427
can not be identified with @xmath421 .
since @xmath257 and @xmath427 can not both be identified with @xmath420 , we have that for @xmath428 or @xmath429 , @xmath424 is identified with @xmath425 .
but then the basic restriction forces the gluing to be as in figure [ marked decagons ] . in case
( e ) , all three cases of figures [ marked decagons ] ( fifth polygon ) and [ marked decagons1 ] are possible . to see that there are no others ,
let @xmath430 be the labelling of the four consecutive edges .
if @xmath257 is identified with @xmath431 and @xmath432 with @xmath421 then the point vertex on @xmath257 on the side opposite to @xmath259 becomes a regular point and we are not in @xmath1 . if @xmath257 is identified with @xmath431 and @xmath432 with @xmath420 then by the basic restriction we must identify @xmath427 with @xmath426 and hence @xmath259 with @xmath421 , which violates the basic restriction for the pair @xmath433 . if @xmath257 is identified with @xmath431 and @xmath432 with @xmath426 then by the basic restriction for the sides
@xmath434 we have the polygon on the left side of figure [ marked decagons1 ] .
so we have dealt with all cases in which @xmath257 is identified with @xmath431 , and by symmetry , when @xmath432 is identified with @xmath421 .
now suppose @xmath257 is identified with @xmath420 and @xmath432 with @xmath426 . by the basic restriction @xmath259 is identified with @xmath431 and @xmath427 with @xmath421 .
then the vertex between @xmath420 and @xmath426 becomes a regular point and we are not in @xmath1 .
the remaining case is @xmath257 identified with @xmath426 and @xmath432 with @xmath420 .
if @xmath259 is identified with @xmath431 we have the polygon on the right side of figure [ marked decagons1 ] and if @xmath259 is identified with @xmath421 then the vertex between @xmath259 and @xmath427 becomes a regular point . in case ( f ) , applying the basic restriction to the pair of consecutive sides @xmath419 , @xmath252 which form a side of the hexagon , we see that we must identify @xmath424 with @xmath425 .
then the discussion reduces to case ( d ) .
[ cor : measure zero ] let @xmath435 or @xmath436 with respect to the measure of theorem [ thm : mcmullen eigenform ] , almost every surface in @xmath20 has no convex presentation . in order to prove that a set of translation surfaces has measure zero , it is enough to show that it has measure zero in the cover of @xmath1 corresponding to translation surfaces equipped with a _ marking _ ,
i.e. a fixed identification of @xmath41 with @xmath384 such that singularities map to @xmath389 . fixing a marking , we note that the possible edges of convex polygons @xmath49 may be labelled by homotopy classes of curves beginning and ending at @xmath389 .
in particular there are countably many ways of using a marked decagon @xmath49 to get a convex presentation . for
each such marked decagon @xmath49 and each surface @xmath41 constructed via @xmath49 with a convex presentation , we have exhibited a neighborhood of @xmath41 in which the set of surfaces for which @xmath49 is convex , is of measure zero .
the result follows using the fact that a countable union of sets of measure zero has measure zero .
[ sec : another question ] theorem [ thm : main h(1,1)](v ) resolves a question of veech ( * ? ? ?
* ; * ? ? ?
* question 5.2 ) .
namely veech asked whether there are non - lattice surfaces with no convex presentations , and as we have seen , almost every surface in @xmath24 has these properties . in case of an affirmative answer , veech asked a more refined question .
namely ( in our notations ) , given @xmath334 , let @xmath96 denote the hyperelliptic component of one of the strata @xmath438 , let @xmath439 denote the subset of surfaces without convex presentations , and let @xmath440 denote the closure of @xmath439
. is there a surface in @xmath439 whose orbit - closure is @xmath440 ?
our analysis shows that in the stratum @xmath1 , the answer to this question is negative .
namely , by theorem [ thm : main h(1,1)](iii , iv ) @xmath439 consists of : therefore @xmath440 consists of @xmath26 , and a finite list of closed @xmath13-orbits . since @xmath369 and @xmath4 are closed and disjoint , part ( a ) of this list shows a negative answer to veech s question in @xmath1 .
we remark that the list in ( b ) is nonempty . in light of proposition
[ prop : interesting square tiled ] , we know that there are at least three @xmath13-orbits of arithmetic surfaces outside @xmath26 without strictly convex presentations .
namely , this will be satisfied by the @xmath13-orbit of any square tiled surface made of @xmath441 squares , which is not in @xmath24 . since 5 and 7
are prime we can use @xmath442 to get two examples in @xmath443 and one in @xmath444 .
a. eskin , m. mirzakhani and a. mohammadi , _ isolation , equidistribution and orbit closures for the @xmath445-action on moduli space _
, preprint ( 2012 ) , available online at http://www.math.uchicago.edu/~eskin/ r. j. simpson , _ convex lattice polygons of minimum area , _ bull .
, * 42(3 ) * ( 1990 ) 353367
. w. a. veech , _ teichmller curves in moduli space , eisenstein series and an application to triangular billiards _ , invent . math . * 97 * ( 1989 ) , no .
3 , 553583 . | we give infinite lists of translations surfaces with no convex presentations .
we classify the surfaces in the stratum @xmath0 which do not have convex presentations , as well as those with no strictly convex presentations .
we show that in @xmath1 , all surfaces in the eigenform loci @xmath2 , @xmath3 or @xmath4 have no strictly convex presentation , and that the list of surfaces with no convex presentations in @xmath5 is finite and consists of square - tiled surfaces .
we prove the existence of non - lattice surfaces without strictly convex presentations in all of the strata @xmath6 .
a _ translation surface _ is a union of polygons with pairs of parallel edges identified by translation , up to cut and paste equivalence .
these structures have been intensively studied in recent years in connection with questions in group theory , geometry , complex analysis , and dynamics we recall the definitions in [ sec : general ] and refer to the recent surveys @xcite for detailed definitions and discussions of work on translation surfaces .
the cut and paste equivalence means that the same translation surface has many presentations as unions of polygons .
it is well - known ( see @xcite ) that a connected translation surface may be presented as a single polygon with an even number of edges identified pairwise .
a natural question is whether there is such a presentation , in which the polygon can be taken to be convex . in his 1992 hayashibara
forum lecture , veech @xcite was the first to exhibit surfaces with no presentations as convex polygons .
we review his examples and give new ones in [ sec : examples ] .
these include the veech double @xmath7-gons , the escher staircases , the ward examples , and the bouw - mller examples .
translation surfaces are naturally grouped in strata , which are moduli spaces of translation surfaces for which the combinatorics of singularities is fixed .
although the constructions in [ sec : examples ] yield infinitely many examples , they only give rise to finitely many in each stratum . in this paper
we take up the question of classifying surfaces with no convex presentations in the simplest strata , namely the genus two strata @xmath8 and @xmath1
. it will be useful to distinguish between _ convex _ and _ strictly convex _ polygons .
although , strictly speaking , polygons are never strictly convex , we will say that a polygon is strictly convex if none of its vertices is in the convex hull of the other vertices ; equivalently , the internal angles at vertices of the polygon are strictly less than @xmath9 . in making this definition
we always assume that polygons have no spurious vertices , in other words , edges of polygons are not strictly contained in larger straight segments on which the identification map of edges extends continuously .
we will see that strictly convex presentations only arise in the hyperelliptic components of the strata @xmath10 and @xmath11 .
let @xmath12 .
there is a natural @xmath13-action on each stratum of translation surfaces , and , as noted by veech @xcite , the property of having no ( strictly ) convex presentations is @xmath13-invariant .
moreover the property of having no strictly convex presentations is closed .
thus the question of classifying surfaces without convex presentations is intimately connected to the question of understanding the closures of @xmath13-orbits .
our results crucially rely on the work of mcmullen @xcite describing the possible orbit closures for the @xmath13-action on @xmath8 and @xmath1 .
we remind the reader that by work of calta @xcite and mcmullen @xcite ( see [ sec : genus 2 ] for a review ) , the @xmath13-orbits of lattice surfaces in @xmath8 are in bijective correspondence with symbols @xmath14 , where @xmath15 is the discriminant , and is congruent to a square mod 4 , and where for @xmath16 there are symbols @xmath17 corresponding to even and odd spin structures . some surfaces in @xmath1 have jacobians which are eigenforms for an action of an order of discriminant @xmath18 in the field @xmath19 .
we will denote the locus of such surfaces in @xmath1 by @xmath20 . in this paper , by an _
arithmetic surface _ we mean a translation cover of a torus branched over one point , i.e. a surface which is in the @xmath13-orbit of a square - tiled surface . note that lattice surfaces which are eigenforms of discriminant @xmath18 , where @xmath18 is a square , are arithmetic .
the following are our main results .
[ thm : main ] up to the @xmath13-action , the list of surfaces in @xmath0 which have no strictly convex presentation are those corresponding to symbols @xmath21 up to the @xmath13-action , the list of surfaces in @xmath8 with no convex presentation are those corresponding to symbols @xmath22 [ thm : main h(1,1 ) ] let @xmath23 .
then : 1 . there is no surface in @xmath24 with a strictly convex presentation .
2 . if @xmath25 then the eigenform locus @xmath20 contains a surface with a strictly convex presentation .
3 .
any surface in @xmath1 which has no strictly convex presentation is either arithmetic or belongs to @xmath26 .
4 .
the number of @xmath13-orbits of arithmetic surfaces with no strictly convex presentations in @xmath27 is finite .
5 .
[ item : answer to veech ] for each @xmath28 , there are surfaces in @xmath20 which are not lattice surfaces and have no convex presentations . the space @xmath24 parametrizes torus covers of low degree , and assertion ( i ) is a special case of a more general statement ( see proposition [ remark : more general ] ) which exhibits loci of torus covers with no strictly convex presentation for any genus @xmath29 .
assertion ( v ) resolves a question of veech ( * ? ? ? * ; * ? ? ?
* question 5.2 ) . using theorem [ thm : main h(1,1 ) ] , in [ sec : another question ] we answer a more refined question of veech .
our proof of ( iii ) is ineffective , and it would be interesting to write down the full list of examples of the arithmetic surfaces as in ( iii ) .
we exhibit two such examples , in discriminant 49 , in
[ subsec : square zoo ] .
we briefly describe the proofs of our results .
clearly a convex polygon contains a strip running through it which connects parallel edges on opposite sides of the polygon .
when @xmath29 , this shows that a translation surface with a convex presentation has a simple cylinder , i.e. a cylinder which is bounded on both sides by a single saddle connection .
the existence of a simple cylinder is a nontrivial restriction and this observation underlies our construction ( in [ sec : examples ] ) of many new examples of surfaces with no convex presentations . in genus two ,
the dynamical results of mcmullen allow us to conclude that a surface with no convex presentation is either a lattice surface or an eigenform . by further results of calta and mcmullen , the core direction of
the simple cylinder is a completely periodic direction . by work of mcmullen ,
calta and bainbridge , completely periodic directions on genus two surfaces are described in terms of prototypes , which are explicit presentations of completely periodic surfaces .
there are finitely many prototypes for each discriminant @xmath18 in @xmath8 , and in @xmath1 , prototypes also depend on two continuous real parameters . in [
sec : nec and suff ] we show that there is a canonical polygonal presentation associated with each prototype , which may or may not be convex .
we call this the _ canonical polygon _ associated with the prototype . a strictly convex presentation , if it exists , must be the canonical polygon of some prototype . to prove theorem [ thm : main ] , we first exhibit explicitly , for all @xmath30 , a particular prototype whose canonical polygon is convex .
for the remaining cases @xmath31 we use a computer program , discussed in [ subsec : computer algorithm ] , to list all prototypes and test their canonical polygon for convexity .
figure [ fig : another figure ] shows two prototypes in @xmath32 and @xmath33 and shows the corresponding canonical polygon . using this program
we obtain the list .
sample output of our computer program is shown in figure [ fig : sample output ] .
appealing to results of hubert , lelivre and mcmullen we find that lattice surfaces which are arithmetic do admit a one - cylinder convex presentation .
we use this to obtain the list .
( 2.75,-0.65 ) + ( 1.25,0 ) ; ( a ) at ( 0,0 ) ; ( b ) at ( 0,-1.73205 ) ; ( c ) at ( 1.73205,-1.73205 ) ; ( d ) at ( 1.73205,0 ) ; ( e ) at ( 3,0 ) ; ( f ) at ( 4,1 ) ; ( g ) at ( 2.73205,1 ) ; ( h ) at ( 1,1 ) ; ( ge ) at ( 2.366,0 ) ; ( eg ) at ( 3.366,1 ) ; ( g1 ) at ( 2.73205,0 . ) ; ( g2 ) at ( 3.73205,1 ) ; ( g3 ) at ( 3.73205,0 ) ; ( g4 ) at ( 0.73205,1 ) ; ( g5 ) at ( 0.73205,0 ) ; ( a ) ( d ) ( h ) cycle ; ( g ) ( ge )
( e ) cycle ; ( e ) ( eg ) ( f ) cycle ; ( h )
( d ) ; ( d ) ( g ) ; ( e ) ( eg ) ; ( ge ) ( g ) ; ( g ) ( e ) ; ( a ) ( d ) ; ( h ) ( 1,0 ) ; ( g ) ( g1 ) ; ( g2 )
( intersection of e f and g2g3 ) ; ( intersection of a h and g4g5 ) ( g5 ) ; ( a ) node (b ) node (c ) node (d ) node (e ) node (f ) node (g ) node (h ) node (0,0 ) cycle ; \(a ) at ( 0,0 ) ; ( b ) at ( 0,-1.73205 ) ; ( c ) at ( 1.73205,-1.73205 ) ; ( d ) at ( 1.73205,0 ) ; ( g ) at ( 0.73205,2 ) ; ( h ) at ( 1,1 ) ; ( hh ) at ( 1,1 ) ; ( hh ) at ( 0.73205,-2.73205 ) ; ( gg ) at ( 1,-3.73205 ) ; ( h1 ) at ( 0,1 ) ; ( hh1 ) at ( 1.73205,-2.73205 ) ; ( hl ) at ( intersection of a g and h1h ) ; ( a ) ( d ) ( h ) cycle ; ( a ) ( h ) ( hl ) cycle ; ( hl ) ( h ) ( g ) cycle ; ( a ) ( h ) ; ( b ) ( c ) ; ( intersection of a
g and h1h ) ( h ) ; ( intersection of gg c and hh1hh )
( hh ) ; ( c ) ( hh ) ; ( a ) ( d ) ; ( h ) ( 1,0 ) ; ( g ) ( 0.73205,0 . ) ; ( a ) node (b ) node (hh ) node ( gg ) node ( c ) node (d ) node (h ) node (g ) node ( 0,0 ) cycle ; ( 2.3,-0.65 ) + ( 1.7,0 ) ; ( a ) at ( 0,0 ) ; ( b ) at ( 0,-1.3028 ) ; ( c ) at ( 1.3028,-1.3028 ) ; ( d ) at ( 1.3028,0 ) ; ( e ) at ( 3,0 ) ; ( f ) at ( 4,1 ) ; ( g ) at ( 2.3028,1 ) ; ( h ) at ( 1,1 ) ; ( ge ) at ( 2.1514,0 ) ; ( eg ) at ( 3.1514,1 ) ; ( g1 ) at ( 2.3028,0 . ) ; ( g2 ) at ( 3.3028,1 ) ; ( g3 ) at ( 3.3028,0 ) ; ( g4 ) at ( 0.3028,1 ) ; ( g5 ) at ( 0.3028,0 ) ; ( a ) ( d ) ( h ) cycle ; ( g ) ( ge )
( e ) cycle ; ( e ) ( eg ) ( f ) cycle ; ( h )
( d ) ; ( d )
( g ) ; ( e ) ( eg ) ; ( ge ) ( g ) ; ( g ) ( e ) ; ( a ) ( d ) ; ( h ) ( 1,0 ) ; ( g ) ( g1 ) ; ( g2 ) ( intersection of e f and g2g3 ) ; ( intersection of a h and g4g5 ) ( g5 ) ; ( a ) node (b ) node (c ) node (d ) node (e ) node (f ) node (g ) node (h ) node (0,0 ) cycle ; \(a ) at ( 0,0 ) ; ( b ) at ( 0,-1.3028 ) ; ( c ) at ( 1.3028,-1.3028 ) ; ( d ) at ( 1.3028,0 ) ; ( g ) at ( 0.3028,2 ) ; ( h ) at ( 1,1 ) ; ( hh ) at ( 1,1 ) ; ( hh ) at ( 0.3028,-2.3028 ) ; ( gg ) at ( 1,-3.3028 ) ; ( h1 ) at ( 0,1 ) ; ( hh1 ) at ( 1.3028,-2.3028 ) ; ( hl ) at ( intersection of a g and h1h ) ; ( a ) ( d ) ( h ) cycle ; ( a ) ( h ) ( hl ) cycle ; ( hl ) ( h ) ( g ) cycle ; ( a ) ( h ) ; ( b ) ( c ) ; ( intersection of a
g and h1h ) ( h ) ; ( intersection of gg c and hh1hh )
( hh ) ; ( c ) ( hh ) ; ( a ) ( d ) ; ( h ) ( 1,0 ) ; ( g ) ( 0.3028,0 . ) ; ( a ) node (b ) node (hh ) node ( gg ) node ( c ) node (d ) node (h ) node (g ) node ( 0,0 ) cycle ; [ fig : another figure ] to prove theorem [ thm : main h(1,1)](i ) we use the fact that surfaces in @xmath24 admit a branched covering map of a twice punctured torus , of low degree , to rule out strictly convex canonical polygons .
note that in @xmath1 the number of prototypes for a fixed @xmath18 is no longer finite so we can not use a brute force computer search as in the case @xmath34 for theorem [ thm : main h(1,1)](ii ) , we discuss separately the discriminants @xmath18 which do or do not appear in the list . if @xmath18 does not appear in , the existence of a strictly convex polygon in @xmath8 forces the existence of a nearby strictly convex polygon of discriminant @xmath18 in @xmath1 .
for the remaining cases of @xmath35 which do appear in we give explicit strictly convex canonical polygons for some prototypes ( these were found by a computer search ) . assertion ( iii ) of theorem
[ thm : main h(1,1 ) ] follows from assertion ( ii ) and mcmullen s dynamical results for the @xmath13-action on @xmath1 .
the proof of the finiteness statement in theorem [ thm : main h(1,1)](iv ) relies on a deep recent finiteness result of @xcite . to prove theorem [ thm : main h(1,1)](v ) we show that convex but not strictly convex presentations may be ` easily destroyed ' by a local perturbation in @xmath20 .
since there are no strictly convex presentations in @xmath36 , this implies that the set of surfaces in @xmath26 with no convex presentations at all must be of full measure , with respect to the natural measure on these loci .
* acknowledgments : * our interest in the question of convex presentations of translation surfaces was inspired by thierry monteil s ` banana peel argument ' @xcite .
we thank matt bainbridge for useful discussions .
we thank pat hooper for permission to use his picture ( figure [ fig : bm ] ) of the bouw - mller surfaces , and for useful discussions about these surfaces .
this research was supported by the anr projet blanc geodym , israel science foundation grant 190/08 and european research council grant dlgaps 279893 . |
the distinguishability of a quantum theory from its classical counterpart is formulated by the heisenberg uncertainty principle @xcite , which bounds our prediction ability for a quantum system . in terms of entropic measures
, this uncertainty relation can be recast in a modern information - theoretical form @xcite , which states that @xmath0 , where @xmath1 and @xmath2 denote the shannon entropy for the probability distribution of the measurement outcomes . since the overlap @xmath3 between observables @xmath4 and
@xmath5 does not depend on specific states to be measured , the right - hand side ( rhs ) of the inequality provides a fixed lower bound and a more general framework of quantifying uncertainty than the standard deviations @xcite .
moreover , once using quantum memory to store information about the measured system , the entropic uncertainty bound could even be violated @xcite .
all these entropic uncertainty relations play an important role in many quantum information processing .
however , constructed from the probability distribution of a measurement taken , entropic function is still a rather coarse way of measuring the uncertainty of a set of measurements ( see ref .
@xcite for a recent review ) .
for instance , one can not distinguish the uncertainty inherent in obtaining any combination of outcomes for different measurements . to overcome
this defect , a new form of uncertainty relation , i.e. , fine - grained uncertainty relation ( fgur ) , has been proposed recently @xcite . for a set of measurements labeled by @xmath6 , associating with every combination of possible outcomes @xmath7
, there exist a set of inequalities @xmath8 where @xmath9 is the probability of choosing a particular measurement , @xmath10 is the probability that one obtains the outcome @xmath11 after performing measurement @xmath6 on the state @xmath12 . to measure the uncertainty ,
the maximum in function @xmath13 should be taken over all states allowed on a particular system .
it can be proved that one can not obtain outcomes with certainty for all measurements simultaneously when @xmath14 . on the other hand , for the so - called maximally certain state ( mcs )
, the inequality ( [ fgur ] ) can be saturated . since its introduction , many applications have been found for the fgur .
for instance , it was shown @xcite that the fgur could be used to discriminate among classical , quantum , and superquantum correlations involving two or three parties . moreover ,
the uncertainty bound in ( [ fgur ] ) could be optimized once the measured system is assisted by a quantum memory @xcite .
dramatically , a profound link between the fgur and the second law of thermodynamics has been found @xcite , which claims that a violation of uncertainty relation implies a violation of thermodynamical law .
other studies from various perspectives could be found in @xcite .
while most of studies on uncertainty relations are nonrelativistic , a complete account of these relations requires one to understand them in relativistic regime , which may even shed new light on quantum gravity @xcite . in the previous work @xcite
, we have shown that , besides the choice on the observers , the entropic uncertainty bound should also depend on the relativistic motion state of the observer who performs the measurement .
this new character of quantum - memory - assisted entropic uncertainty relation ( eur ) is a direct result of entanglement generation in a relativistic system . in this paper
, we explore the fgur under the decoherence rooting in the relativistic motion of quantum systems , and find the uncertainty bound does depend on the motion state of the system .
we first consider observer undergoes an _ uniform _ acceleration relative to an inertial reference .
since two frames differ in their description of a given quantum state due to the unruh effect , the concept of measurement becomes observer - dependent , which implies a nontrivial relativistic modification to the fgur . without loss of generality , beyond the single - mode approximation
, we show that the unruh effect modifies the uncertainty bound for the measurements in @xmath15 mutually unbiased bases ( mubs ) , which are intimately related to complementarity principle .
dramatically , we find that , for a noninertial observer , the measurements in general mubs could be distinguished from each other , while they share same uncertainty bound in inertial frame .
we extend these results to an alternative scenario , where , to prevent the unruh decoherence , an observer is restricted in a single rigid cavity which undergoes an _ nonuniform _ acceleration .
we show that the uncertainty could be degraded by the nonuniform acceleration of cavity during particular epoch , while the uncertainty bound itself exhibits a periodic evolution with respect to the duration of the acceleration .
this phenomenon can be attributed to the entanglement generation between the field modes in single cavity that plays the role of quantum memory . moreover , except the acceleration - duration time with integer periods , the measurements in different mubs are distinguishable by the corresponding uncertainty bounds .
we first investigate the fgur for an observer with uniform acceleration @xmath16 , who performs projective measurements on the quantum state constructed from free field modes . for the noninertial observer traveling in , e.g. , right rindler wedge , field modes in left rindler wedge are unaccessible . the information loss associated with the acceleration horizon results in a thermal bath . from quantum information point of view @xcite ,
this celebrated unruh effect induces a nontrivial evolution of quantum entanglement between field modes which plays a prominent role in most quantum - information protocols . for simplicity ,
considering fermionic field with few degrees , the most general vacuum state @xmath17 should be annihilated by unruh operators @xmath18 @xcite c_k , u&=&q_rc_k , r^+q_lc_k , l , q_r^2+q_l^2=1 , + c_k , r&=&_k ,-
d^_k , , + c_k , l&=&_k ,- d^_k , . where @xmath19 and @xmath20 are real parameters , and @xmath21 .
the particle and antiparticle operators @xmath22 and @xmath23 in respective rindler wedge @xmath24 satisfy the usual anti commutation relations . by analytic construction to whole spacetime , the proper unruh modes are symmetric between rindler wedges i and ii .
for particular frequency @xmath25 , the unruh vacuum is @xcite & & + ( |1100-|0011 ) [ unruh - f-0 ] and the first excitation is & & + q_l(|1101+|0001 ) [ unruh - f - one ] where we introduce the notations anti - particle vacua are denoted by @xmath26 and @xmath27 .
it should be noted that different operator ordering in fermonic systems could lead to nonunique results in quantum information @xcite .
for instance , if we rearrange operator ordering in ( [ notation ] ) as @xmath28 , then a new fock basis is defined @xmath29 .
this so - called physical ordering @xcite , in which all region i operators appear to the left of all region ii operators , was proposed to guarantee the entanglement behavior of above states would yield physical results .
hereafter , we adopt this particular operator ordering . to explore how the relativistic motion of an observer could influence the fgur
, we consider a scenario in which the state to be measured should be prepared in an inertial frame .
after that , the observer undergoes an uniform acceleration and performs measurements . since the state should be described in the corresponding rindler frame , information would lose via unruh effect .
therefore , the uncertainty obtained by the accelerated observer would be motion - dependent .
we illustrate above insight by measurements @xmath30 and @xmath31 , behaving as the best measurement basis , where pauli operators @xmath30 and @xmath31 with equal probability @xmath32 are chosen @xcite .
remarkably , along with @xmath33 , three set of their eigenvectors form the mubs in hilbert space with dimension @xmath15 , which plays central role to theoretical investigations and practical exploitations of complementarity properties @xcite . for all pure states @xmath34 , @xmath35,\phi\in[0,2\pi)$ ] , the corresponding density matrix @xmath36 should be rewritten according to the transformation ( [ unruh - f-0 ] ) and ( [ unruh - f - one ] ) . since the field modes in rindler wedge ii is unaccessible to observer , after tracing over the modes in wedge ii ,
the reduced density matrix is _ red&&_|| + & = & |00_00|(^2c^4+^2q_l^2c^2 ) + & & + |11_11|(^2s^4+^2q_r^2s^2 ) + & & + |01_01|(^2s^2c^2 ) + & & + |10_10|(^2s^2c^2+^2q_r^2c^2+^2q_l^2s^2 ) + & & -|00_01|e^iq_lsc^2 + & & + |00_10|e^-iq_rc^3 + & & -|00_11|^2q_rq_lsc + & & + |01_11|e^-iq_rs^2c + & & + |10_11|e^iq_ls^3+(h.c . ) _ with abbreviation @xmath37 . after performing the projective measurements @xmath30 and @xmath31 on particle sector , we have the probabilities for the outcomes @xmath38 p(0^z|_z)_&&(|0^+_0|_red)=c^2(^2+^2q_l^2 ) , + p(0^x|_x)_&&(|+^+_+|_red ) + & = & + q_rc . here ,
for later convenience , we chose the measurements in basis @xmath39 and @xmath40 which are the eigenstates of pauli matrix @xmath30 and @xmath31 restricted in rindler wedge i. therefore , lhs of fgur ( [ fgur ] ) should be u&&[p(0^z|_z)_+p(0^x|_x ) _ ] + & = & [ q_rc+(q_r^2+q_l^2 + 1)c^2 + 1 ] for a particular unruh mode with fixed acceleration , one can estimate that the maximum of @xmath41 which is @xmath42 should always be obtained with @xmath43 and @xmath44 @xcite .
this means that the mcs which saturates the inequality ( [ fgur ] ) is independent of the acceleration of the observer , which indicates that once we choose the bases enabling an optimal uncertainty on average in an inertial frame , the corresponding measurements should maintain their optimality for all noninertial observer
. this remarkable phenomenon could be useful in many real quantum process , for instance , the bb84 states @xmath45 and @xmath46 in quantum cryptography @xcite . explicitly , the dependence of the fine - grained uncertainty bound with outcomes @xmath38 on the acceleration parameter and choice of unruh modes could be expressed as _ ( 0^x,0^z)&=&[c^2(1+q_l^2 + q_r^2)+q_rc+1 ] [ bound1 ] above calculation can extend to any other pairs of outcomes @xmath47 , @xmath48 and @xmath49 , which all give the same bound @xmath50 in inertial frame @xcite .
dramatically , we find that the nontrivial unruh effect could distinguish these four pairs of measurements and separate them into two categories .
for instance , we have p(1^z|_z)_&&(|1_1|_red ) + & = & ^2s^2+^2(q_r^2+q_l^2s^2 ) , + p(1^x|_x)_&&(|-_-|_red ) + & = & -q_rc . which give _ ( 1^x,1^z)&=&[(1+c^2)q_r^2+(1+q_l^2)s^2+q_rc+1 ] + [ bound2 ] for the mcs with @xmath51 and @xmath44 .
after straightforward calculations , it can be shown that the uncertainty bound @xmath52 is also dependent on the direction of @xmath53axis , that measurements @xmath47 share the same bound ( [ bound2 ] ) with @xmath49 , while @xmath48 has the same bound ( [ bound1 ] ) with @xmath38 .
this is definitely a new feature of fine - grained uncertainty bound triggered by noninertial relativistic motion . to further explore the influence of unruh effect on fugr
, we illustrate the uncertainty bounds ( [ bound1 ] ) and ( [ bound2 ] ) for three different choice of unruh modes , as depicted in fig .
[ fgur1 ] .
is dependent on the acceleration parameter @xmath54 and choice of unruh modes .
three set of curves correspond to the choice of unruh modes with @xmath55 ( black solid ) , @xmath56 ( red dashed ) and @xmath57 ( blue dashed - double - dotted ) . ] for the case with @xmath55 , where the minkowskian annihilation operator is taken to be one of the right or left moving unruh modes , the noninertial observer would detect a single - mode state once the field is in a special superposition of minkowski monochromatic modes from an inertial perspective @xcite . under this single - mode approximation ( sma ) ,
commonly assumed in the old literature on relativistic quantum information @xcite , we recover the standard result @xmath50 for vanishing acceleration .
as @xmath54 growing , the value of @xmath52 decreases , indicating an increment on measurement uncertainty . while there is no essential difference in the behavior of measurement uncertainty for outcomes @xmath38 and @xmath48 with various @xmath20 , however , the uncertainty bounds ( [ bound2 ] ) are sensitive with the choice of unruh modes . surprisingly ,
as @xmath54 growing , for the unruh modes with large @xmath20 , we observe a decrement on measurement uncertainty for outcomes @xmath47 and @xmath49 , which means that unruh effect could even lift the uncertainty bound in ( [ fgur ] ) . finally , as illustrated in fig .
[ fgur1 ] , we find that the distinguishability between the measurements in mubs is a common feature for any choice of unruh modes . to explain this , recall that , by definition , a set of orthonormal bases @xmath58 for a hilbert space @xmath59 where @xmath60 is called unbiased iff @xmath61 , @xmath62 holds for all basis vectors @xmath63 and @xmath64 belong to different bases , i.e. , @xmath65 . from an inertial perspective , the mubs are intimately related to complementarity principle @xcite , which indicates that the measurement of a observable reveals no information about the outcome of another one if their corresponding bases are mutually unbiased .
however , for a noninertial observer , the bases @xmath58 should be transformed according to proper bogoliubov transformations , which in general breaks the orthonormality . in other word , the mubs in inertial frame would become non - mubs from a noninertial perspective . therefore , for the observer undergoing an uniform acceleration , we conclude that unruh effect could distinguish measurements in mubs .
we now discuss an alternative scenario in which observer is localized in a rigid cavity .
while the rigid boundaries of the cavity protect the inside observer from the unruh effect , the relativistic motion of the cavity would still affect the entanglement between the free field modes inside @xcite , therefore leading to a motion - dependent uncertainty bound @xcite . for simplicity
, we consider a @xmath66-dimensional model , where the cavity with length @xmath67 imposes the dirichlet conditions on the eigenfunctions @xmath68 of the hamiltonian .
a typical trajectory of nonuniform - moving cavity contains three segments referred as ( i ) when the cavity maintains its inertial status , then ( ii ) begins to accelerate at @xmath69 , following the killing vector @xmath70 , and finally ( iii ) the acceleration ends at rindler time @xmath71 , and the duration of the acceleration in proper time measured at the center of the cavity is @xmath72 .
the dirac field can be expanded in quantized eigenfunctions as @xmath73 in segment i , and similarly be expressed by @xmath74 in segment ii and @xmath75 in segment iii. the nonvanishing anticommutators @xmath76 define the vacuum @xmath77 .
any two field modes in distinct regions can be related by bogoliubov transformations like @xmath78 and @xmath79 , where the coefficients can be calculated perturbatively in the limit of small cavity acceleration @xcite .
more specifically , by introducing the dimensionless parameter @xmath80 , which is the product of the cavity s length and the acceleration at the center of the cavity , the coefficients can be expanded in a maclaurin series to @xmath81 order , @xmath82 , and similarly @xmath83 .
we start from a pure state @xmath84 in segment i. after the uniform acceleration , we can express this state in segment iii by means of the bogoliubov transformations , which contains modes within all frequency . throughout the process
, we assume that the observer can only be sensitive to modes in particular frequency @xmath85 .
therefore all other modes with frequency @xmath86 should be traced out in the density matrix @xmath87 , which leads to _
red&=&_k|_k_k| + & = & |_k_k|(-f^-_k+f^+_k ) + & & + |_k^++_k|(+f^-_k - f^+_k ) + & & + |_k^+_k|e^-i(g_k+^(2)_kk ) + & & + |_k^+_k|e^i(g_k+^(2)_kk)^ * where the coefficients are @xmath88 and @xmath89 .
the probability of measurements @xmath38 are p(0^z|_z)_&&(|_k_k|_red ) + & = & -f^-_k+f^+_k , + p(0^x|_x)_&&(|++|_red ) + & = & \{1+[e^-i(g_k+^(2)_kk)]}. the uncertainty bound should be the maximum of lhs of fgur ( [ fgur ] ) , @xmath90 $ ] . along the analysis before , we know that the acceleration of cavity would not change the mcs with parameters @xmath43 and @xmath44 .
therefore , we obtain the uncertainty bound for the cavity system _ ( 0^x,0^z)=[2+(1-f_+)+f_-+(g_k+^(2)_kk ) ] + [ fgur3 ] the coefficients in the bound has been given in @xcite , which are f_+&=&_p=-^|e_1^k - p-1|^2|a^(1)_kp|^2 + & = & [ 4(k+s)^2(q_6(1)-q_6(e_1))+q_4(1)-q_4(e_1)]and @xcite f_- & & f^+_k - f^-_k=(_p0-_p<0)|e_1^k - p-1|^2|a^(1)_kp|^2 + & = & 2(k+s)[q_5(1)-q_5(e_1)]+p(k , s , e_1 ) with @xmath91 characterizing the self - adjoint extension of the hamiltonian .
here we use the notation @xmath92 $ ] , li is the polylogarithm and @xmath93 .
@xmath94 is a polynomial summing for all terms with odd number @xmath95\ ] ] and ( g_k+^(2)_kk)&=&1-h^2\{(+ ) + & & - } in previous section , we show that an uniformly - accelerating observer can distinguish the measurements in mubs which share the same uncertainty bounds in an inertial frame .
here we generalize this to the scenario with rigid cavity . to proceed ,
we calculate the probabilities of measurements @xmath49 p(1^z|_z)_&&(|_k^++_k|_red ) + & = & + f^-_k - f^+_k , + p(1^x|_x)_&&(|++|_red ) + & = & \{1-[e^-i(g_k+^(2)_kk)]}. which give the uncertainty bound @xmath96 for mcs with @xmath97 _ ( 1^x,1^z)=[2+(1-f_+)-f_-+(g_k+^(2)_kk ) ] + [ fgur4 ] we depict above uncertainty bounds for measurements performed within cavity in fig .
[ fgur3 ] .
we find that both uncertainty bounds ( [ fgur3 ] ) and ( [ fgur4 ] ) are now periodic in time @xmath98 , which measures the duration of the cavity acceleration , with the period @xmath99 .
the evolution of the uncertainty bound can be interpreted as a result of the entanglement generation between the field modes in the single rigid cavity that plays the role of quantum memory @xcite . by properly choosing the parameters to ensure that @xmath100 with @xmath101 , the uncertainty bounds are protected @xcite , recovering the value @xmath102 as in inertial case . while the uncertainty change is very small due to the low acceleration approximation @xmath103 imposed
, our results could provide a novel way to detect the relativistic effect by future quantum metrology .
depends on the duration time of acceleration of rigid cavity .
we choose @xmath104 . for each pair of measurements ( @xmath38 and
@xmath48 , @xmath49 and @xmath47 ) , three cures from top to bottom correspond to parameters @xmath105 .
the parameter @xmath106 $ ] characterizes the duration time of cavity acceleration . to demonstrate the low acceleration approximation ,
the uncertainty is estimated under @xmath107 . ] on the other hand , for arbitrary acceleration duration @xmath108 , measurements with outcomes @xmath38 and @xmath49 can be distinguished from each other by the corresponding uncertainty bounds ( [ fgur3 ] ) and ( [ fgur4 ] ) . by a straightforward calculation
, it can be shown that measurements @xmath47 share the same bound ( [ fgur4 ] ) with @xmath49 , while @xmath48 has the same bound ( [ fgur3 ] ) with @xmath38 .
therefore , same as unruh effect , we conclude that the relativistic motion of a rigid cavity can provoke the distinguishability between the measurements in mubs that share the same bound @xmath102 for an inertial observer .
in this paper , we have explorer the nontrivial relativistic modification to the fgur .
we have shown that , for the observer undergoes a large acceleration , the associated unruh effect could increase or reduce the fine - grained uncertainty bounds , depending on the choice of unruh modes .
nevertheless , the mcs still could be independent of the acceleration , which indicates that once we choose the bases enabling an optimal uncertainty on average in an inertial frame , the corresponding measurements should maintain their optimality for all noninertial observer .
moreover , we have shown that the measurements in mubs , sharing same uncertainty bound in inertial frame , could be distinguished from each other when the observer undergoes a nonvanishing acceleration . in an alternative scenario
, we have investigated the fgur for the measurements restricted in a single rigid cavity , where the uncertainty bound itself exhibits a periodic evolution with respect to the duration of the acceleration
. this phenomenon could be understand by the entanglement generation in a single rigid cavity that plays the role of quantum memory .
our results provide a novel way to investigate the relativistic effect from a quantum - information perspective , that could be experimental tested by quantum metrology @xcite . our results could be linked to many interesting issues .
for instance , we can generalize above analysis to fundamental mutually unbaised bases in higher dimensional hilbert spaces @xcite , where a @xmath109-qubit can be truncated from free scalar field modes with infinite levels . on the other hand ,
the fascinating link between fgur and the second thermodynamical law has been explored in @xcite , which proved that a deviation of the fgur implies a violation of the second law of thermodynamics . in this spirit , by investigating the influence of the relativistic motion of observer on a thermodynamical cycle , one could relate the relativistic effect to thermodynamics in an information - theoretic way .
in particular , one can investigate the uncertainty relations in general dynamical spacetimes @xcite , e.g. , cosmological background , where the entanglement generation under the evolution of spacetime is expected to play a dramatical role in measurements @xcite . with a proper designed thermodynamical cycle
, such investigation might provide us a new perspective on quantum gravity .
this work is supported by the australian research council through dp110103434 .
j. f. thanks li - hang ren for stimulating discussions .
h. f. acknowledges the support of nsfc , 973 program through 2010cb922904 . | one of the most important features of quantum theory is the uncertainty principle .
amount various uncertainty relations , the profound fine - grained uncertainty relation ( fgur ) is used to distinguish the uncertainty inherent in obtaining any combination of outcomes for different measurements . in this paper
, we explore this uncertainty relation in relativistic regime
. for observer undergoes an uniform acceleration who immersed in an unruh thermal bath , we show that the uncertainty bound is dependent on the acceleration parameter and choice of unruh modes .
dramatically , we find that the measurements in mutually unbiased bases ( mubs ) , sharing same uncertainty bound in inertial frame , could be distinguished from each other for a noninertial observer . on the other hand ,
once the unruh decoherence is prevented by utilizing the cavity , the entanglement could be generated from nonuniform motion .
we show that , for the observer restricted in a single rigid cavity , the uncertainty exhibits a periodic evolution with respect to the duration of the acceleration and the uncertainty bounds can be degraded by the entanglement generation during particular epoch . with properly chosen cavity parameters , the uncertainty bounds could be protected .
otherwise , the measurements in different mubs could be distinguished due to the relativistic motion of cavity .
implications of our results for gravitation and thermodynamics are discussed . |
a maximally extended and regular covering of the schwarzschild vacuum is now a fundamental part of any introduction to general relativity . almost always the covering is given by way of the implicit kruskal @xcite - szekeres @xcite procedure @xcite .
however , as emphasized long ago by ehlers @xcite , an explicit maximally extended and regular covering of the schwarzschild vacuum is known and was first given by israel @xcite . unfortunately , despite the fact that these coordinates offer many advantages over the kruskal - szekeres coordinates , they are almost never used . in this paper
i extend israel s procedure to the schwarzschild - de sitter class of vacua in four dimensions and then to arbitrary dimensions @xmath0 assuming a @xmath1 sphere as in the tangherlini generalization of the schwarzschild vacuum .
moreover , it is made clear that these coordinates offer important advantages over the kruskal - szekeres procedure .
these advantages include : an explicit representation of the line element that can be extended to arbitrary dimension , a simultaneous covering of both the black hole and cosmological horizons and derivation by direct integration of einstein s equations @xcite without recourse to coordinate transformations .
we start with a spherically symmetric spacetime in coordinates @xmath2 where @xmath3 is a null four - vector so that the line element takes the form @xcite @xmath4 where @xmath5 is the metric of a unit two - sphere ( @xmath6 ) .
further , setting @xmath7 ( so that trajectories of constant @xmath8 and @xmath9 are radial null geodesics affinely parameterized by @xmath10 ) it follows that @xmath11 . defining @xmath12 and rewriting @xmath13 as @xmath14 we set @xmath15 in ( [ generalmetric ] ) which remains a completely general spherically symmetric spacetime .
the expansion of @xmath16 reduces to @xmath17 associated with the vector field @xmath18 we have @xmath19 , @xmath20 and @xmath21 the trajectories of constant @xmath22 and @xmath9 are radial null geodesics ( affinely parameterized by @xmath14 ) only for @xmath23 .
note that if @xmath24 then @xmath25 is tangent to a radial null geodesic and clearly @xmath10 is again affine .
the associated geodesics immediately follow as @xmath26 where @xmath27 is a constant .
important examples of this special case are given below .
for the spacetime ( [ generalmetric ] ) with @xmath15 , @xmath29 and @xmath30 where @xmath31 ( @xmath32 ) and @xmath33 ( @xmath34 ) are constants , it follows that @xmath35 @xmath36 and @xmath37 where @xmath38 @xmath39 is the ricci tensor , @xmath40 the ricciscalar and @xmath41 the weyl tensor . viewing @xmath28 as a constant of nature and @xmath33 a property of the vacuum , it is clear from ( [ lambda ] ) that the specification of @xmath31 does not determine the physical situation uniquely . since , by ( [ r(u , w ) ] ) , the axes are centered on @xmath42 , it is important to distinguish ranges in @xmath31 . writing @xmath43 with @xmath44 and @xmath45 it follows immediately from ( [ lambda ] ) that @xmath46 it follows from ( [ crange ] ) that all values of @xmath47 are determined uniquely by @xmath48 using the following ranges : @xmath49 for @xmath50 and @xmath51 for @xmath52 .
the latter range is of no interest and is not discussed here .
the one independent invariant derivable from the riemmann tensor without differentiation can be taken to be ( [ weylscalar ] ) and so with @xmath14 and @xmath10 extending over the reals , the vacua are maximally extended and regular for @xmath53 ( and for all finite @xmath54 if @xmath55 ) . in particular , note that @xmath56 with ( [ r(u , w ) ] ) it follows that @xmath57 from ( [ metric ] ) and ( [ r(u , w ) ] ) it follows that trajectories of constant @xmath22 and @xmath9 are radial null geodesics only for @xmath58 .
more generally , the radial null geodesics that satisfy @xmath59 with @xmath60 given by ( [ metric ] ) and ( [ r(u , w ) ] ) , can be written down in terms of elementary functions ( as explained below ) .
note that for these trajectories @xmath61 as @xmath62 ( @xmath63 ) and @xmath64 as @xmath65 .
we distinguish the cases : @xmath55 ( de sitter ) , @xmath66 ( bertotti - kasner ) , @xmath67 ( schwarzschild ) , @xmath68 ( schwarzschild - de sitter ) , @xmath69 ( schwarzschild - anti de sitter ) and discuss them below . for the case @xmath67 the coordinates were first obtained by israel @xcite . not covered as special cases are anti de sitter space ( see appendix a ) and the degenerate schwarzschild - de sitter spacetime ( see appendix b ) . with @xmath55 , @xmath70 , @xmath71 and we have de sitter space .
the metric simplifies to @xmath72 trajectories with four tangents @xmath73 are radial null geodesics ( so @xmath10 is affine for both radial null directions ) .
we can write these geodesics in the form @xmath74 where @xmath27 is a constant . the negative cosmological horizons ( @xmath75 ) are then given by @xmath76 where the expansion @xmath77 ( @xmath78 ) vanishes .
some details are shown in fig .
[ desitterfig ] . with @xmath79
we have bertotti - kasner space .
the metric simplifies to @xmath80 trajectories with four tangents @xmath81 are radial null geodesics ( so again @xmath10 is affine for both radial null directions ) but now @xmath82 .
we can write these geodesics in the form @xmath83 where @xmath27 is a constant .
the @xmath84 plane is like that of de sitter space now with @xmath85 ( again @xmath76 ) distinguishing the branches of the @xmath86 geodesics .
this space shows that the birkhoff theorem does not extend directly to @xmath87 ( see the case @xmath88 below ) @xcite . with @xmath67 , @xmath89 , @xmath90 and
we have the schwarzschild vacuum in israel coordinates @xcite .
the metric simplifies to @xmath91 trajectories with four tangents @xmath92 are radial null geodesics so @xmath10 is not affine .
now @xmath93 .
we write these geodesics in the form @xmath94 where @xmath27 is a constant .
some details are shown in fig .
[ schwarzschild ] .
again writing @xmath43 it follows that for @xmath95 there is another constant @xmath96 for which @xmath97 .
the constant is given by @xmath98 with @xmath99 the trajectories @xmath100 are radial null geodesics ( with non - zero expansion ) tangent to the cosmological " horizons .
note that @xmath101 as @xmath102 and @xmath103 as @xmath104 .
now define @xmath105 and @xmath106 trajectories with four tangents @xmath107 where @xmath108 and @xmath109 where @xmath27 is a constant , are radial null geodesics .
now we find that the expansion is given by @xmath110 where @xmath111 the choice @xmath112 reproduces the bertotti - kasner result and the choice @xmath113 reproduces the schwarzschild vacuum result both for @xmath114 as given above . for @xmath95 these geodesics can be given explicitly in terms of elementary functions and are reproduced in appendix c. some details are shown in fig .
[ schwarzdesitter ] .
for @xmath116 there are no cosmological horizons but equations ( [ adef ] ) through ( [ fdef ] ) hold as given above . again
the associated radial null geodesics can be given explicitly in terms of elementary functions and are reproduced in appendix d. some details are shown in fig . [ antidesitterimage ] .
this section parallels ii but extends the construction to arbitrary dimensions ( @xmath118 ) . write @xmath119 where @xmath120 is the metric of a unit @xmath1 sphere .
again @xmath3 is tangent to null geodesics affinely parameterized by @xmath10 but now the associated expansion is given by @xmath121 further , associated with the vector field @xmath18 we again have @xmath19 and @xmath122 but now the associated expansion is @xmath123 the trajectories for which all coordinates are constant except @xmath14 are radial null geodesics ( affinely parameterized by @xmath14 ) only for @xmath23 . finally , as before , if @xmath24 then @xmath124 is tangent to a radial null geodesic with @xmath10 again affine and the associated geodesics immediately follow as @xmath125 where @xmath27 is a constant .
this section generalizes section iii above . for dimensions
@xmath126 consider the spaces described by @xmath127 where the function @xmath128 is given by @xmath129 and @xmath54 signifies the function @xmath130 with @xmath31 ( @xmath32 ) and @xmath33 ( @xmath34 ) constants .
it follows that @xcite @xmath131 @xmath132 and @xmath133 where @xmath134 again , the one independent invariant derivable from the riemmann tensor without differentiation can be taken to be ( [ weylscalargen ] ) and so with @xmath14 and @xmath10 extending over the reals , the spaces described by ( [ metricsgen ] ) are maximally extended and regular for @xmath53 ( and for all @xmath54 if @xmath135 , see below ) .
in particular , note that @xmath136 again viewing @xmath28 as a constant of nature and @xmath33 a property of the vacuum , the specification of @xmath31 does not determine the physical situation uniquely .
this is discussed below . in view of the generality of the cases considered ,
the formulae given are remarkably simple .
consider the hyper - spherically symmetric spacetime @xmath137 note that @xmath54 is now a coordinate .
it is not difficult to show that the unique form of @xmath60 which satisfies ( [ riccigen ] ) is given by @xmath138 where @xmath139 is a constant ( which we take @xmath140 ) .
this solution was apparently first discussed by tangherlini @xcite .
the coordinates are , of course , defective at @xmath141 , but the associated roots depend on @xmath0 and are not so easy to find from ( [ curvaturef ] ) .
the previous construction circumvents this unnecessary algebra as follows : for ( [ curvature ] ) with ( [ curvaturef ] ) we find @xmath142 and so from ( [ weylscalargen ] ) and ( [ weylsqcurv ] ) ( assuming , without loss in generality , the same coordinates for @xmath120)we have the relation @xmath143 which gives @xmath144 only for @xmath145 . now substituting for @xmath139 from ( [ mrelation ] ) into ( [ curvaturef ] ) and writing an associated root to @xmath141 as @xmath42 we obtain ( [ lambdagen ] ) . in this way
we never need to deal with ( [ curvaturef ] ) and so all dimensions @xmath118 are equally difficult to deal with as regards the location of horizons .
the previous section raises the issue of mass " . in spacetime with @xmath89 the concept of mass in the spherically symmetric case
is well established @xcite .
the geometrical mass in @xmath146 dimensions for spaces admitting a @xmath147sphere has been defined previously by way of the sectional curvature @xcite . whereas this is merely a formal definition , it is of interest to compare this definition with @xmath33 and @xmath139 . in dimension @xmath0
define @xmath148 in the special case @xmath149 it follows that @xmath150 more generally , for @xmath151 , it follows that @xmath152 where @xmath28 is given by ( [ lambdagen ] ) .
we can , of course , replace @xmath139 by @xmath33 in ( [ geomasslambda ] ) by solving for @xmath33 from ( [ mrelation ] ) .
we now consider some specific cases .
we define hyper - spherical de sitter space by the requirement @xmath135 so that from ( [ weylscalargen ] ) and ( [ weylsqcurv ] ) we have @xmath153 and so @xmath154 the metric follows as @xmath155 trajectories with four tangents @xmath156 are radial null geodesics ( so @xmath10 is affine for both radial null directions ) .
we can write these geodesics in the form @xmath157 where @xmath27 is a constant .
the negative cosmological horizons ( @xmath158 ) are then given by @xmath76 where the expansion @xmath77 ( @xmath159 ) vanishes .
the associated diagram is qualitatively the same as fig .
[ desitterfig ] .
with @xmath160 we have hyper - spherical bertotti - kasner space .
the metric simplifies to @xmath161 trajectories with four tangents @xmath162 are radial null geodesics ( so again @xmath10 is affine for both radial null directions ) and again @xmath82 .
we can also write these geodesics in the form ( [ bertottigeo ] ) .
asymptotically flat static vacuum black holes ( admitting the @xmath1 sphere ) are unique @xcite and given by the tangherlini generalization of the schwarzschild vacuum @xcite . in our notation
these correspond to the case @xmath67 .
global , regular and explicit coordinates for these spaces have been given previously @xcite .
some further discussion is given in appendix e. more generally , the radial null geodesics that satisfy @xmath163 with @xmath128 given by ( [ fgen ] ) and @xmath54 by ( [ rfngen ] ) , are of interest . for these trajectories , excluding the cases discussed above , it is clear that @xmath61 as @xmath164 and @xmath64 as @xmath165 ( @xmath166 ) . from ( [ lambdagen ] ) , writing @xmath43 we have @xmath167 . in the range
@xmath168 there is another constant @xmath169 for which @xmath170 .
the trajectories @xmath171 are radial null geodesics and are tangent to the cosmological " horizons .
it is also clear that for @xmath172 the solutions to ( [ othernullhigh ] ) evolve in a fundamentally different way as in the case @xmath145 .
in fact , the qualitative behavior of all solutions to ( [ othernullhigh ] ) can be obtained without explicit integration .
the essential conclusion is that the figures given for @xmath145 hold , qualitatively , for @xmath173 .
this is discussed in detail elsewhere @xcite .
a maximally extended , explicit and regular covering of the schwarzschild - de sitter vacua in arbitrary dimension ( @xmath126 ) has been given .
it has bee stressed that these coordinates offer important advantages over the kruskal - szekeres procedure that include : an explicit representation of the line element that can be extended to arbitrary dimension , a simultaneous covering of both the black hole and cosmological horizons and derivation by direct integration of einstein s equations without recourse to coordinate transformations . in view of the generality of the problem solved ,
the resultant formulae obtained are of a remarkably simple form .
part of this work was reported at the the dark side of extra dimensions " workshop held in may 2005 at the banff international research station ( birs ) , banff , alberta .
it is a pleasure to thank werner israel and don page for comments and taylor binnington for drawing the figures .
this work was supported by a grant from the natural sciences and engineering research council of canada and was made possible by use of _ grtensoriii _ @xcite .
anti - de sitter space is not contained in ( [ metric ] ) with ( [ r(u , w ) ] ) but can be given in the form ( [ generalmetric ] ) with @xmath15 and is included here for completeness .
it is given by @xmath174 where the constant @xmath175 .
equations ( [ ricci ] ) and ( [ ricciscalar ] ) of course hold and the space is conformally flat
. trajectories with four tangents @xmath176 are radial null geodesics so @xmath10 is again affine .
we can write these geodesics in the form @xmath177 where @xmath27 is a constant . with @xmath14 and @xmath10 extending over the reals , ( [ antidesitter ] ) is maximally extended and regular .
the degenerate schwarzschild - de sitter black hole has @xmath178 like the bertotti - kasner space but it is not contained in ( [ metric ] ) with ( [ r(u , w ) ] ) but can be given in the form ( [ generalmetric ] ) with @xmath15 and is included for completeness .
coordinates for degenerate black holes are seldom discussed @xcite .
now @xmath179 with @xmath180 in addition to the radial null geodesics given by constant @xmath14 and @xmath58 , the other radial null geodesics can be given by @xmath181 where @xmath182 is the lambert w function @xcite with @xmath183 and @xmath184 where @xmath27 is a constant .
some details are shown in fig .
[ degenschwarzdesitterimage ] .
for @xmath95 the radial null geodesics with tangents @xmath114 can be given in a simple form : @xmath185 is a constant of motion where @xmath186 and @xmath187 writing @xmath188 with @xmath189 the radial null geodesics with tangents @xmath114 can be given in the form @xmath190 where @xmath27 is a constant .
in the present notation we set @xmath67 and so have ( [ tangherlini ] ) with @xmath191 where @xmath192 with @xmath145 we recover ( [ schw ] ) .
even implicit forms of @xmath54 in the kruskal - szekeres procedure are not known for for all @xmath0 ( _ e.g _ @xmath193
@xcite ) . however here
we need only substitute for @xmath0 and insert ( [ rfn ] ) into ( [ tangherlinif ] ) to give a maximally extended , explicit and regular coverings of the space .
for example , with @xmath194 we have @xmath195 solutions to ( [ othernull ] ) with ( [ tangherlinif ] ) and ( [ rfn ] ) can be given explicitly in terms of elementary functions in some cases .
for example , with @xmath194 we have m. d. kruskal , phys .
119 * , 1743 ( 1960 ) .
g. szekeres , publ .
debrecen * 7 * , 285 ( 1960 ) ( now available at general relativity and gravitation * 34 * , 2001 , ( 2002 ) ) .
although usually obtained by way of coordinate transformations , the kruskal - szekeres coordinates can be obtained in an algorithmic way without guessing transformations .
see , for example , k. lake in _ vth brazilian school of cosmology and gravitation _ edited by m. novello ( world scientific , singapore , 1987 ) .
j. ehlers in _ relativity , astrophysics and cosmology .
proceedings of the summer school held , 14 - 26 august , 1972 at the banff centre , banff , alberta _ edited by w. israel ( reidel , boston , 1973 ) .
w. israel , phys . rev . * 143 * , 1016 ( 1966 ) .
israel coordinates have been obtained independetly by d. w. pajerski and e. t. newman , j. math .
* 12 * , 1929 ( 1971 ) and by t. klsch and t. strobl , class .
* 13 * , 1191 ( 1996 ) arxiv:(gr - qc/9507011 ) .
israel @xcite ( in the published form of his work ) introduced the coordinates by way of a transformation which , in his notation reads as @xmath197 . in this paper einstein s equations for vacua
have been solved directly without prior coordinates .
this is not possible for kruskal - szekeres coordinates wherein even at @xmath145 with @xmath198 the vacuum field equations are difficult to solve .
see , for example , lake @xcite .
we use einstein s equations ( with cosmological constant ) , geometrical units and a signature of @xmath1 .
the sign conventions for @xmath10 and @xmath14 is that @xmath10 can decrease and @xmath14 increase only in spacelike directions . for convenience , explicit functional dependence
is usually shown only on the first appearance of a function . but see k. lake , phys .
d * 20 * , 370 ( 1979 ) , j. podolsky , gen .
. grav . * 31 * , 1703 ( 1999 ) arxiv : gr - qc/9910029 , s. gao , phys .
d * 68 * , 044028 ( 2003 ) arxiv:(gr - qc/0207029 ) and r. balbinot , s. fagnocchi , a. fabbri , s. farese and j. navarro - salas , phys .
d * 70 * , 064031 ( 2004 ) arxiv:(hep - th/0405263 ) for a detailed discussion see w. rindler , physics letters a * 245 * , 363 ( 1998 ) .
this is defined by @xmath199 .
see , for example , r. m. corless , g. h. gonnet , d. e. g. hare , d. j. jeffrey , and d. e. knuth , advances in computational mathematics * 5 * , 329 ( 1996 ) . if the cosmological constant @xmath200 is defined ( in dimension @xmath0 ) by the vacuum field equations @xmath201 where @xmath202 is the @xmath0 dimensional einstein tensor , then @xmath203 .
f. r. tangherlini , nuovo cimento * 27 * , 636 ( 1963 ) .
see , for example , w. c. hernandez and c. w. misner , astrophys .
j. * 143 * , 452 ( 1966 ) , m. e. cahill and g. c. mcvittie , j. math .
phys , 11 , * 1360 * ( 1970 ) , e. poisson and w. israel , phys .
rev d * 41 * , 1796 ( 1990 ) , t. zannias , phys .
d * 41 * , 3252 ( 1990 ) and s. hayward , phys .
d * 53 * , 1938 ( 1996 ) arxiv:(gr - qc/9408002 ) k. lake , arxiv:(gr - qc/0606005 ) s. hwang , geometriae dedicata * 71 * , 5 ( 1998 ) , g. w. gibbons , d. ida and t. shiromizu , prog .
phys . suppl . * 148 * , 284 ( 2003 ) arxiv:(gr - qc/0203004 ) .
k. lake , j. cosmol .
jcap * 10 * 007 ( 2003 ) arxiv:(gr - qc/0306073 ) k. lake ( in preparation ) .
this is a package which runs within maple .
it is entirely distinct from packages distributed with maple and must be obtained independently .
the grtensorii software and documentation is distributed freely on the world - wide - web from the address _ http://grtensor.org _
grtensoriii software is in development . | maximally extended , explicit and regular coverings of the schwarzschild - de sitter family of vacua are given , first in spacetime ( generalizing a result due to israel ) and then for all dimensions @xmath0 ( assuming a @xmath1 sphere ) .
it is shown that these coordinates offer important advantages over the well known kruskal - szekeres procedure . |
in his nowadays classical studies on the kinetic theory of gases , the danish physicist m. knudsen experimentally observed that , no matter how an inert gas was injected into a pipeline , the direction in which a molecule rebounds from the pipeline s solid wall is asymptotically independent of its initial trajectory .
that is , the fraction of particles leaving the surface in a given direction is proportional to @xmath0 , where @xmath1 is the angle that such a particle s trajectory defines , measured with respect to the normal to the surface ( @xcite ) .
this behaviour is referred to as the * _ knudsen s ( cosine ) law _ * ever since . in the experiment ,
the gas is injected at a very low pressure so that interactions between particles are negligible .
the physical justification of knudsen s law is the following .
first of all , one assumes that particles bounce at the pipeline s wall _ elastically_.
this means that the energy of a particle is preserved in a collision which , in turn , implies the _ refection law _ : the angle that the incident direction forms with the normal to a surface , @xmath2 , equals the angle @xmath3 of the reflected direction , @xmath4 .
it has been proved that this law is valid in a first approximation if we do not take into account thermal effects . since we also assume that particles do not interact with one another
, we must conclude that the microscopic irregularities on the pipeline s surface are responsible for the destruction of any particular pattern in the original gas distribution . indeed , even if we assume that the bounces are perfectly elastic , microscopic holes in the boundary of the pipeline and imperfections in relief are dimensionally comparable to the molecules of the gas and have therefore a disruptive ( i.e. , unpredictable ) effect on particle collisions : after many bounces , we are in the so - called _
knudsen s regime _ in which the reflected direction is independent of the incident one .
however , this argument does not explain by itself why the reflected angles are distributed according to the
_ cosine law_. the theory of _ billiards _ helps clarify this point ( @xcite , @xcite ) .
the irregularities of the pipeline surface can be reasonably modelled as cavities or microscopic cells with a dispersive geometry such that , once a particle has entered one of them , it comes out of it with a rather arbitrary direction , even if all the collisions inside that cavity are elastic . if the pipeline is made of a uniform material , we can model one such cell as a billiard table and the pipeline wall as an infinite row of such billiard tables that a particle moving freely enters and exits . with this description , knudsen s cosine law
can be then seen as a consequence of the fact that , for sufficiently dispersive billiard geometries , the liouville measure is the unique measure preserved by the billiard flow ( see @xcite for more details ) .
unfortunately , this need not be the case for polygonal ( i.e. , non - dispersive ) tables such as fig . [ fig pipeline ] .
so polygonal geometries require a slightly different approach . since the characteristic size of this cell is infinitely small compared with the diameter of the pipeline , the exact position along the open side of the cell ( dotted line in fig .
[ fig pipeline ] ) at which the particle enters a cell coming from the previous one is considered to be randomly distributed with uniform probability . following @xcite , we will refer to these billiards as _
random billiards_. more concretely , the dynamics of a particle bouncing inside the pipeline are given by the _ first return map _ to the open side of a the billiard cell , which defines a markov process .
thus , the dynamics are characterised by a transition operator @xmath5 such that , for any reflected direction @xmath6 $ ] , @xmath7 is the probability that a molecule takes a direction in @xmath8 $ ] after the next rebound .
these concepts will be reviewed in sections [ section preliminaries ] and [ section random billiards ] . in this formalism
, knudsen s law can be mathematically written as follows .
let @xmath9 be the initial angle distribution with which the gas is injected into the pipeline .
that is , if @xmath8 $ ] , @xmath10 represents the proportion of particles that will hit the pipeline s wall for the first time with an incident angle @xmath11 .
denote by @xmath12 the distribution after @xmath13 collisions . in this context
, knudsen s cosine law can be expressed in two different ways .
the * _ strong knudsen s law _ * claims that , for almost any initial distribution @xmath9 , after many collisions @xmath12 converges to the cosine distribution , @xmath14 . \label{eq strong law}\ ] ] on the other hand , the * _ weak knudsen s law _ * states that if , after any collision , we count the number of particles reflected with a given angle @xmath1 and then divide by the total number of particles , this quantity is proportional to @xmath15 .
explicitly,@xmath16 obviously ( [ eq strong law ] ) implies ( [ eq weak law ] ) , so the strong knudsen s law implies the weak one .
deterministic dynamical systems can be sometimes understood as idealizations of _ real _ systems , the latter usually subjected to negligible perturbations in a first approximation .
the use of a dynamical approach to study a gas is one such example . in this paper
, we will study the properties of the random billiard obtained as a result of modelling the rough surface of the pipeline by a zig - zag geometry as shown in fig .
[ fig pipeline ] .
our billiard cell is an isosceles triangle with open side @xmath17 defined by an angle @xmath18 .
this model was introduced in @xcite but has not been studied in depth yet . for example , in @xcite nothing is said about the ergodic properties of this billiard table when @xmath19 ( it is shown that it is not ergodic under the liouville measure if @xmath20 though ) . thus it is not at all clear if the dynamics resulting from a random billiard as in fig .
[ fig billiard table ] will in fact follow knudsen s law ( weak or strong ) .
in this paper , we will prove that it is actually the case .
more concretely , the main contributions of our paper are the following : 1 .
little is known about the ergodicity of the flow of deterministic billiards given by irrational polygon tables ( @xcite ) .
remarkably , we show that , for a concrete example , randomising the billiard in a sensible way ( i.e. , choosing the point at which the particle enters the billiard randomly but uniformly distributed on one side of the table ) implies its ergodicity and exactness with respect to the liouville measure . 2 . unlike other attempts to prove ( weak ) knudsen s law , where one usually assumes very irregular geometries at a microscopic level responsible for the dispersive effects
, we deal with an extremely simple pattern .
this has additional advantages as a simpler model capturing the main features of a system can be simulated and developed more easily .
3 . in the literature ,
ergodicity of random billiards relies on the ergodicity of the first return map ( see @xcite ) .
we use here a different approach and take advantage of a skew - type representation of random maps recently introduced in @xcite .
the techniques used to prove the exactness of our model can be certainly adapted to study other ( random ) billiard tables and random maps in general .
finally , we prove that the strong knudsen s law holds for our model . as we mentioned before
, even if one assumes dispersive billiard geometries , one will only prove that the first return map is ergodic , i.e. , the weak law .
the paper is structured as follows : in section [ section model ] we introduce the random billiard behind our model .
we recall some basic definitions and properties of billiards in section [ section preliminaries ] and the concept of random billiard in section [ section random billiards ] . in section [ section random maps ]
, we review the skew - type representation of random maps introduced in @xcite and show the relationship between this representation and the dynamical evolution of absolutely continuous measures .
we prove that this skew - type representation is exact in section [ section exactness of s ] , which uses some auxiliary results presented separately in section [ section ergodicity of ss ] .
finally , in section [ section knudsen s law ] , we illustrate our results with numerical simulations .
suppose we have a gas of non - interacting particles moving freely inside a pipeline .
our model consists of a two dimensional infinite pipeline whose rough walls are modelled as a sequence of cells built as a juxtaposition of a _ fundamental cell _ ( fig .
[ fig billiard table ] ) . in this section
, we are going to fix the geometry of such cell and introduce the main hypothesis behind the dynamics of the particles .
the content of this section is extracted from @xcite .
both the geometry and dynamics of our model are summarised in figure [ fig pipeline ] .
we assume that the walls of the pipeline are not smooth but describe a serrated regular pattern built from a fixed isosceles triangle with an open side .
we characterise that triangle by one of its angles @xmath21 .
the dynamics of a particle moving freely inside the pipeline are as follows : the particle enters one of the cells with some angle @xmath2 and bounces on its sides elastically , that is , according to the reflection law .
the particle eventually comes out of the cell with a given direction @xmath22 , then crosses the pipeline , and reaches another cell on the opposite wall , and the process recurs . since the diameter of the pipeline is several orders of magnitude bigger than the characteristic length @xmath17 of the cell , it is plausible to think that , every time the particle enters a new cell through the open side , it does so at a point uniformly distributed on @xmath17 . in the literature ,
one encounters examples of particle dynamics on bounded domains where the random perturbation is introduced in the observations of @xmath22 every time the particles bounces off the wall ( see for example @xcite ) , which implies that collisions are no longer elastic .
this perturbation is explained as a consequence of the roughness of the wall .
observe that , in our model , randomness is introduced in a completely different way .
we assume elastic collisions but , instead of considering dispersive geometries responsible for the random bounces , the noise is introduced in a sensible way without modifying the ( rather elementary ) geometry of the problem . the trajectory that a particle describes after it comes out of a billiard cell until it enters another one on the opposite wall is completely irrelevant for dynamical purposes .
in practice , we can better understand our dynamical system as the closed billiard table obtained by fixing together two triangular cells so that they share the same open side as in figure [ fig billiard table ] .
when the particle crosses that open side , it retains its direction but it is assigned a different position on @xmath17 according to a uniform law . at a uniformly distributed random point.,width=377,height=181 ] the * _ incoming angle _ * @xmath2 ( resp . *
_ outgoing angle _ * ) is then the angle in @xmath23 $ ] with which the particle enters ( resp .
leaves ) the cell measured with respect to the normal to the open side @xmath17 .
one often shifts @xmath23 $ ] by @xmath24 and measure @xmath1 in @xmath25 $ ] , i.e. with respect to horizontal . in @xcite , for any incoming angle @xmath26 $ ] , feres gives the different possible directions with which a particle may come out of the fundamental cell ( fig .
[ fig billiard table ] ) .
they are four , given by maps @xmath27\rightarrow\mathbb{r}$ ] defined as@xmath28 and any @xmath29 applies with certain probability @xmath30 .
the probabilities associated to the maps ( [ eq tau maps ] ) are as follows:@xmath31{ll}1 & \theta\in\lbrack0,\alpha)\\ u_{\alpha}\left ( \theta\right ) & \theta\in\lbrack\alpha,\pi-3\alpha)\\ 2\cos(2\alpha)u_{2\alpha}(\theta ) & \theta\in\lbrack\pi-3\alpha,\pi-2\alpha)\\ 0 & \theta\in\lbrack\pi-2\alpha,\pi ] \end{array } \right .
\smallskip\nonumber\\ p_{2}\left ( \theta\right ) & = \left\ { \begin{array } [ c]{ll}0 & \theta\in\lbrack0,\pi-3\alpha)\\ u_{\alpha}\left ( \theta\right ) -2\cos(2\alpha)u_{2\alpha}(\theta ) & \theta\in\lbrack\pi-3\alpha,\pi-2\alpha)\\ u_{\alpha}\left ( \theta\right ) & \theta\in\lbrack\pi-2\alpha,\pi-\alpha)\\ 0 & \theta\in\lbrack\pi-\alpha,\pi ] \end{array } \right .
\smallskip\nonumber\\ p_{3}\left ( \theta\right ) & = \left\ { \begin{array } [ c]{ll}0 & \theta\in\lbrack0,2\alpha)\\ 2\cos(2\alpha)u_{2\alpha}(-\theta ) & \theta\in\lbrack2\alpha,3\alpha)\\ u_{\alpha}\left ( -\theta\right ) & \theta\in\lbrack3\alpha,\pi-\alpha)\\ 1 & \theta\in\lbrack\pi-\alpha,\pi ] \end{array } \right .
\smallskip\nonumber\\ p_{4}\left ( \theta\right ) & = \left\ { \begin{array } [ c]{ll}0 & \theta\in\lbrack0,\alpha)\\ u_{\alpha}\left ( -\theta\right ) & \theta\in\lbrack\alpha,2\alpha)\\ u_{\alpha}\left ( -\theta\right ) -2\cos(2\alpha)u_{2\alpha}(-\theta ) & \theta\in\lbrack2\alpha,3\alpha)\\ 0 & \theta\in\lbrack3\alpha,\pi ] , \end{array } \right .
\tag{m2}\label{eq probabilities model}\ ] ] where@xmath32 and @xmath18 is assumed to be smaller than @xmath33 .
as we will see later , this family of maps and probabilities is all we need to prove knudsen s law .
this section aims at recalling the main concepts of the theory of ( deterministic ) billiards and one of its key results , namely , that the _ billiard map _ preserves the liouville measure .
the content of this section was extracted from @xcite , which the reader is referred to for an exhaustive exposition on billiards . a * _ billiard table _ * @xmath34 is the closure of a bounded open domain @xmath35 whose boundary @xmath36 is a finite union of smooth compact curves ( @xmath37 , @xmath38 ) .
@xmath39 are called the * _ walls _ * of the billiard table and are assumed to intersect each other only at their endpoints or * _ corners_*. for any @xmath40 , @xmath41 is defined by a @xmath37 map @xmath42\subset \mathbb{r}\rightarrow\mathbb{r}^{2}$ ] such that the second derivative @xmath43 either vanishes or is identically zero .
this condition prevents the wall @xmath41 from having inflection points or line segments . a wall such that @xmath44 is called * _ focusing _ * or * _ dispersing _ * if @xmath45 points inwards or outwards @xmath34 , respectively .
otherwise @xmath46 and the wall is called * _ flat_*. in this framework , we consider the dynamics of a point particle moving freely within @xmath34 .
let @xmath47 denote the position of a free particle moving within a billiard table .
we suppose that the particle bounces at @xmath48 * _ elastically_*. that is , if it collides with the wall @xmath49 at a point @xmath50 that is not a corner , the incident angle @xmath2 that the velocity @xmath51 forms with the normal vector @xmath13 to @xmath49 at @xmath50 equals the angle of reflection @xmath3 between @xmath13 and @xmath51 after the collision , i.e. @xmath4 .
this implies that @xmath51 has constant norm , which we will assume equal to @xmath52 for the sake of simplicity . under these assumptions , the * _ billiard flow _ * @xmath53 is the flow that this particle defines on the * _ phase space _ * @xmath54 . since the particle follows a straight trajectory between two consecutive collisions , the dynamical information of the system is contained in the geometry of the boundary @xmath48 and how the particle bounces off it .
therefore , instead of studying the billiard flow , first one introduces the cross - section @xmath55 $ ] as the set of all postcollisional velocity vectors , where @xmath56 $ ] measures the angle between @xmath51 and the normal vector @xmath13 pointing to the interior of @xmath34 , and then considers the first return map @xmath57 that the billiard flow induces on @xmath58 .
that is , given a point of @xmath58 , @xmath57 gives the next position at where the particle collides and its velocity after that collision .
the map @xmath57 is often called the * _ billiard map_*. it is a well known result in the theory of billiards that the billiard map preserves the * _ liouville measure _ * @xmath59 on @xmath58 , where @xmath60 is the lebesgue measure on @xmath48 and @xmath61 is the measure on @xmath23 $ ] given by @xmath62 , @xmath63\right ) $ ] ( ( * ? ? ?
* ; * ? ? ?
* lemma 2.35 ) ) .
however , except for a few general results , little is known about the ergodicity of billiard maps with respect to the liouville measure for explicit examples of boundary geometries .
for example , dispersive billiards are ergodic ( ( * ? ? ?
* ; * ? ? ?
* theorem 6.20 ) ) while regular polygons are not .
finally , as we pointed out in section [ section model ] , it is customary in the literature to translate @xmath23 $ ] and measure @xmath1 in @xmath64 $ ] , in which case @xmath61 is given by @xmath65 , @xmath66\right ) $ ] .
we will follow this convention throughout the paper .
in this section we are going to get back to the concept of _ random billiard _ and its different representations .
let @xmath34 be a billiard table with boundary @xmath67 .
remove one of the walls of @xmath48 , for example @xmath68 , so that we obtain a billiard table with an _
open side_. without loss of generality , we can assume that the open side is just a line segment @xmath17 as in figure [ fig billiard table ] .
this identification can be carried out by the map @xmath69\subset\mathbb{r}\rightarrow\mathbb{r}^{2}$ ] defining @xmath68 . on the other hand ,
suppose a particle enters the billiard from the open side at @xmath70 with an angle @xmath71 $ ] with respect to @xmath68 .
denote by @xmath72 $ ] the angle associated to the velocity of that particle when it returns to the open side for the first time .
the * _ first return map _ * ( to the open side ) is then the map @xmath73\rightarrow \overline{pq}\times\lbrack0,\pi]$ ] induced from the billiard map @xmath57 such that @xmath74 , i.e. , the particle leaves the billiard at @xmath75 with an angle @xmath76 .
the liouville measure @xmath59 on @xmath77 $ ] induces a measure on @xmath78 $ ] by restriction in a natural way .
we will continue referring to this measure as the liouville measure and denoting it by @xmath59 , where @xmath60 now stands for the lebesgue measure on @xmath17 .
it can be proved that the first return map @xmath79 also preserves the liouville measure ( ( * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* exercise i.5.1 ) ) .
furthermore , [ lemma ergodicity first return]if the billiard map @xmath80\rightarrow\partial d\times\lbrack0,\pi]$ ] is ergodic with respect to @xmath59 , so is the first return map @xmath81\rightarrow\overline{pq}\times\lbrack0,\pi]$ ] .
let @xmath82)$ ] be an invariant set , i.e. , @xmath83 .
this set can be naturally regarded as a measurable set of @xmath77 $ ] since @xmath84 is identified with the wall @xmath68 . since @xmath57 is ergodic , @xmath85 has full @xmath86}$]-measure , which means that@xmath87}(({\textstyle\bigcup\nolimits_{n\geq1 } } f^{-n}\left ( a\right ) ) \bigcap\left ( \overline{pq}\times\lbrack 0,\pi]\right ) ) = 1.\ ] ] if @xmath88\right ) $ ] , then @xmath89 for some @xmath90 and , from the very definition of the first return map that , @xmath91 for some @xmath92 .
therefore,@xmath93\right ) = \bigcup \nolimits_{m\geq1}t^{-m}\left ( a\right ) = a,\ ] ] where the last equality follows from the invariance of @xmath94 . consequently ,
@xmath95}(a)=1 $ ] and @xmath79 is ergodic .
let @xmath96 be the probability space built upon @xmath17 with the normalised lebesgue measure @xmath60 on the borel @xmath97-algebra @xmath98 .
the * _ random billiard _ * associated to @xmath34 is then the time - discrete random dynamical system @xmath99\rightarrow \lbrack0,\pi]~|~x\in\overline{pq}\right\ } $ ] , where @xmath17 is regarded as a probability space .
roughly speaking , the random billiard thus built models the outgoing state a free particle that entered the billiard through the open side at a point uniformly distributed along @xmath17 . when regarded as a probability space , @xmath17 will be denoted by @xmath100 .
the random billiard @xmath99\rightarrow\lbrack0,\pi ] ~|~x\in\omega\right\ } $ ] defines a transition probability kernel @xmath101\times\mathcal{b}\left ( [ 0,\pi]\right ) \rightarrow\lbrack0,1]$ ] given by@xmath102\text { and } a\in\mathcal{b}\left ( [ 0,\pi]\right ) .
\label{eq probability kernel}\ ] ] we say that @xmath9 is * _ invariant _ * with respect to @xmath103\rightarrow\lbrack0,\pi]~|~x\in\omega\right\ } $ ] if @xmath104}k(\theta , a)d\nu(\theta)\text { for any } a\in \mathcal{b}\left ( [ 0,\pi]\right)\ ] ] and that the random billiard is * _ ergodic _ * with respect to @xmath9 if @xmath79 is ergodic with respect to @xmath105 .
for example , @xmath106 , @xmath107\right ) $ ] , is an invariant measure because @xmath59 is invariant with respect to the first return map
@xmath79 ( ( * ? ?
? * ; * ? ? ?
* ; * ? ? ?
* proposition 2.1 ) ) .
moreover , since dispersive billiards are ergodic with respect to the liouville measure , the corresponding random billiard is also ergodic with respect to @xmath61 ( lemma [ lemma ergodicity first return ] ) . in this situation , birkhoff s ergodic theorem implies that @xmath108 where @xmath109\rightarrow\lbrack0,\pi]$ ] is the projection onto the second factor .
that is , the average number of particles reflected with some angle @xmath11 converges to @xmath110 so ( weak ) knudsen s law holds for dispersive billiards . in general , for non - dispersive billiards , little is know about the ergodicity of the first return map .
this is indeed the case of the random billiard introduced in section [ section model ] ( fig .
[ fig billiard table ] ) . recall that , as far as we know , determining whether a general irrational triangular billiard is ergodic is still an open problem ( ( * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* section 7 ) ) and only a few particular examples were proved to be ergodic ( @xcite ) . therefore , at this point , it is not clear at all if knudsen s law holds for the pipeline model introduced in section [ section model ] .
to prove it , the first return map must be replaced with a more convenient representation .
given a random billiard , the probability kernel ( [ eq probability kernel ] ) contains all the dynamical information of the system and determines its properties .
this implies that , as far as the dynamics is concerned , the underlying probability space @xmath100 ( which in section [ section random billiards ] was equal to @xmath96 ) plays a secondary role .
in particular , other probability spaces @xmath111 giving rise to the same probability kernel are available .
indeed , if @xmath112 $ ] is a transition probability kernel on a measurable space @xmath113 , kolmogorov s existence theorem on markov processes guarantees that there exists a probability @xmath114 defined on the borel @xmath97-algebra of the topological space @xmath115 , @xmath116 for all @xmath117 , such that the chain @xmath118 is markovian and has transition probability kernel @xmath119 ( ( * ? ? ?
* ; * ? ? ?
* theorem 2.11 ) ) , where @xmath120 is the projection onto the @xmath121th factor . in this statement
, @xmath122 has to be a @xmath97-compact hausdorff space and @xmath123 the borel @xmath97-algebra .
unfortunately , @xmath124 is not very manageable to work with explicitly .
for instance , the representation @xmath73\rightarrow\overline { pq}\times\lbrack0,\pi]$ ] given by the first return map introduced in section [ section random billiards ] only involves finite dimensional spaces and hence seems more convenient .
we are now going to consider another skew - type representation for random maps recently introduced in @xcite . as we will show in section [ section exactness of s ] , this representation will be crucial to prove the asymptotic properties of the random map defined by our random billiard fig .
[ fig billiard table ] .
let @xmath125 be a general measure space and let @xmath126 be the unit interval regarded as a probability space , where @xmath60 stands for the lebesgue measure .
later on , we will apply the results of this section to @xmath127 $ ] . for any @xmath128 ,
let @xmath129 and @xmath130 $ ] be measurable mappings such that @xmath131 is a measurable partition of the unity , i.e. , @xmath132 for any @xmath133
( compare to ( [ eq probabilities model ] ) and ( [ eq tau maps ] ) in section [ section model ] ) .
we define the * _ random dynamical system _ * @xmath134 such that @xmath135 with probability @xmath136 .
the transition probability kernel of @xmath137 is given by@xmath138 this probability kernel defines the evolution of an initial distribution ( probability measure ) @xmath9 on @xmath113 under the random map @xmath137 iteratively as@xmath139}k(\theta , a)d\nu^{(n)}(\theta),~~n\geq1,\label{eq 15}\ ] ] where @xmath140 .
a measure @xmath9 is called * _ _ * invariant if @xmath141 .
following @xcite , consider now @xmath142 and set @xmath143 such that @xmath144 is a finite measurable partition of @xmath100 .
we define the * _ skew - type representation _ * of the random map @xmath137 as the map @xmath145 such that@xmath146 where@xmath147 the map @xmath148 is well defined because if @xmath149 then @xmath150 necessarily .
moreover , @xmath151 thus defined is @xmath152-measurable ( ( * ? ? ?
* lemma 3.1 ) ) and it is a skew - type representation of the random dynamical system @xmath137 . as we will see , the map @xmath151 is extremely useful to study the properties of @xmath137 .
for example , @xmath9 is an invariant measure of @xmath137 on @xmath153 if and only if @xmath105 is an invariant measure of @xmath151 on @xmath154 ( ( * ? ? ?
* lemma 3.2 ) ) .
moreover , since we know that the sine @xmath65 is invariant by the transition probability kernel ( [ eq 16 ] ) of our billiard table , we conclude that @xmath61 is invariant by @xmath137 built from ( [ eq probabilities model ] ) and ( [ eq tau maps ] ) , and therefore @xmath59 is invariant by the corresponding map @xmath151 . in an abuse of terminology , we will continue calling @xmath59 the * _ liouville measure when referring to _ * @xmath151 . in this context
, we will say that the random map @xmath137 with initial distribution @xmath9 is * _ ergodic _ * ( resp .
* _ mixing _ * , * _ exact _ * ) if @xmath151 is ergodic ( resp . mixing , exact ) with respect to @xmath105 . in section [ section exactness of s ] , we are going to prove that the @xmath151 associated to ( [ eq probabilities model ] ) and ( [ eq tau maps ] ) is exact .
[ h ] s.jpg unlike the first return map @xmath79 ( section [ section random billiards ] ) , the map @xmath151 has no dynamical interpretation and is a purely auxiliary tool to represent and study @xmath137 .
therefore , we need to establish the relationship between @xmath151 and the transition probability kernel ( [ eq 16 ] ) , which carries all the dynamical information .
this is the content of theorem [ theorem 1 ] . before stating that relationship
, we need to introduce some notation and an auxiliary lemma , whose proof is included separately in the appendix for the sake of a clearer exposition .
[ definition slices]we define iteratively the sets @xmath155 , where @xmath156 for any @xmath157 , as@xmath158 [ lemma slices go to zero]let @xmath159 the fibre through @xmath133 . if @xmath160 , then@xmath161 [ theorem 1]let @xmath9 be a measure on @xmath113 and @xmath140 .
then,@xmath162 : = \int_{[0,1)\times x}\mathbf{1}_{a}\left ( \pi_{2}\circ s^{n}\right ) d\left ( \lambda\otimes\nu\right ) = \nu^{(n)}\left ( a\right ) .\ ] ] first of all , it is not difficult to check that@xmath163 this can be proved iteratively rewriting @xmath119 in terms of ( auxiliary ) dirac deltas , @xmath164 where @xmath165 , @xmath166 . on the other hand ,
since @xmath167 is the disjoint union of the sets @xmath168 , @xmath156,@xmath169 where , in the second line , we have used that @xmath170 on the set @xmath168 . if now we apply fubini s theorem,@xmath171 but by lemma [ lemma slices go to zero ] , for any fixed @xmath133,@xmath172
therefore,@xmath173 [ corollary koopman]let @xmath174 , @xmath175 , be the * koopman operator * associated to @xmath151 , i.e. , @xmath176 . let @xmath177 be the projection onto the second factor .
if @xmath9 is a measure on @xmath153 , then@xmath178}u_{s}^{n}\left ( 1_{\pi_{2}^{-1}(a)}\right ) d\left ( \lambda\otimes\nu\right ) , ~a\in \mathcal{b},~n\in\mathbb{n}.\ ] ] let @xmath151 be the skew - type representation of the random billiard introduced in section [ section model ] and let @xmath179 be a probability measure absolutely continuous with respect to the liouville measure @xmath61 with radon - nikodym derivative @xmath180,\mu\right ) $ ] .
one can easily check that @xmath181 , where @xmath182 and @xmath183 $ ] . if @xmath151 is mixing ( or exact ) then , from corollary [ corollary koopman ] , @xmath184 ( ( * ? ? ?
* proposition 4.4.1 ( b ) ) ) , which implies the strong knudsen s law .
we will prove that @xmath151 is exact in section [ section exactness of s ] and give more details about the strong law in section [ section knudsen s law ] .
the proof uses the fact that the pull - back @xmath185 of a characteristic function @xmath186 , @xmath187 \right ) $ ] , can not be invariant by @xmath188 ( section [ section ergodicity of ss ] ) .
it is worth observing that the strong law , unlike the standard approach to random billiards available in the literature from the first return map ( see [ eq 18 ] ) , is a consequence of the asymptotic properties of the skew - type representation @xmath151 .
let @xmath151 be the skew - type representation of the random map ( [ eq tau maps ] ) and ( [ eq probabilities model ] ) . in this section
, we are going to show that @xmath190\rightarrow\lbrack0,1)\times \lbrack0,\pi]$ ] can not leave invariant any characteristic function @xmath191 , where @xmath192 \right ) $ ] has probability @xmath193 and @xmath177 denotes the projection onto the second factor .
this result will be used in the next section to prove that @xmath151 is exact .
first , we will show that our model exhibits a very useful symmetry .
[ prop symmetry]let @xmath151 be the skew - type representation of the random map ( [ eq tau maps ] ) and ( [ eq probabilities model ] ) .
let @xmath194\rightarrow\mathbb{r}$ ] be a function such that @xmath195 is @xmath189-invariant , i.e.,@xmath196 and let @xmath197\rightarrow\lbrack0,\pi]$ ] be defined by @xmath198 .
then @xmath199 is also @xmath189-invariant .
it is not difficult to realise from ( [ eq probabilities model ] ) that@xmath200 just using that the the function @xmath201 introduced in ( [ eq definition function u ] ) has period @xmath202 .
furthermore , straightforward computations shows that@xmath203 given an index @xmath204 , we define its conjugate index @xmath205 in the following manner:@xmath206 on the other hand , ( [ eq 13 ] ) is equivalent to@xmath207 we have to check that this property holds for @xmath208 .
so let @xmath209 $ ] such that @xmath210 .
since @xmath211 and @xmath212 for any @xmath213 , we have that @xmath214 implies@xmath215{r}p_{i}\left ( x\right ) = p_{\widetilde{i}}\circ\phi\left ( x\right ) > 0\\ p_{j}\left ( \tau_{i}\left ( x\right ) \right ) = p_{\widetilde{j}}\circ \phi\circ\tau_{i}\left ( x\right ) = p_{\widetilde{j}}\circ\tau_{_{\widetilde { j}}}\circ\phi\left ( x\right ) > 0 \end{array } \right\ } \longrightarrow(p_{\widetilde{i}}\cdot(p_{\widetilde{j}}\circ \tau_{_{\widetilde{j}}}))\circ\phi\left ( x\right ) > 0.\ ] ] therefore , using the invariance of @xmath216 expressed in ( [ eq 34])@xmath217 so @xmath208 also satisfies ( [ eq 34 ] ) and is @xmath189-invariant .
[ prop g invariant s2 and phi]let @xmath218,\mu\right ) $ ] and suppose that @xmath195 is invariant by @xmath189 and @xmath219 .
if @xmath18 in ( [ eq probabilities model ] ) and ( [ eq tau maps ] ) is irrational , then @xmath216 is constant .
let @xmath220\right/ _ { 0\sim\pi}$ ] be the quotient space obtained by identifying @xmath221 .
this space is homeomorphic to the unit circle @xmath222 .
the function @xmath223\rightarrow\mathbb{r}$ ] induces a function @xmath224:\left .
[ 0,\pi]\right/ _ { 0\sim\pi}-\{[0]\}\rightarrow \mathbb{r}$ ] on the quotient that is well defined except at @xmath225 $ ] , the equivalence class of @xmath226 , because maybe @xmath227 .
we are going to prove that @xmath224 $ ] is invariant by an irrational rotation .
since constants are the only integrable functions invariant by irrational rotations , @xmath224 $ ] must be constant and @xmath216 too .
recall that @xmath228 reads as@xmath229 we are going split the interval @xmath230 $ ] into subintervals and , for each of them , we will explore the invariance properties of @xmath216 implied by ( [ eq 34 ] ) .
( i ) : : if @xmath231 , then @xmath232 so@xmath233 ( ii ) : : if @xmath234 , then @xmath235 . writing @xmath236 with @xmath237,@xmath238 now , since @xmath216 is @xmath219-invariant,@xmath239 therefore,@xmath240 \left ( \left [ x\right ] \right ) = \left [ g\right ] \left ( \left [ x+6\alpha\right ] \right ) \text { if } x\in\lbrack\pi -4\alpha,\pi-3\alpha).\ ] ] ( iii ) : : if @xmath241 then @xmath242 . writing @xmath243 with @xmath237,@xmath244 using @xmath245,@xmath246 that is,@xmath240 \left ( \left [ x\right ] \right ) = \left [ g\right ] \left ( \left [ x+6\alpha\right ] \right ) \text { if } x\in\lbrack\pi -3\alpha,\pi-2\alpha).\ ] ] ( iv ) : : suppose @xmath247 , @xmath248 with @xmath237 . since @xmath249 on @xmath250 , by ( [ eq 34]),@xmath251 since @xmath216 is invariant by @xmath219,@xmath252 summarising , @xmath253 which implies@xmath254\left ( [ x]\right ) = [ g]\left ( [ x+2\alpha]\right ) \text {
if } x\in\lbrack\pi-2\alpha,\pi-\alpha).\ ] ] now @xmath237 and we can use the @xmath189-invariance expressed in the first item so that@xmath240 \left ( \left [ x\right ] \right ) = \left [ g\right ] \left ( \left [ x+2\alpha\right ] \right ) = \left [ g\right ] \left ( \left [ x+6\alpha\right ] \right ) .\ ] ] ( v ) : : finally , let @xmath255 , @xmath256 with @xmath237 . since @xmath257 on @xmath258@xmath259 by ( [ eq 34 ] ) . using the @xmath219-invariance of @xmath216,@xmath260 that is , @xmath261 , which again implies@xmath254\left ( [ x]\right ) = [ g]\left ( [ x+2\alpha]\right ) \text { if } x\in\lbrack\pi-\alpha,\pi).\ ] ] now @xmath262 and @xmath263 as by assumption @xmath264 ; we are in the situation of ( i ) so@xmath240 \left ( \left [ x\right ] \right ) = \left [ g\right ] \left ( \left [ x+2\alpha\right ] \right ) = \left [ g\right ] \left ( \left [ x+6\alpha\right ] \right ) .\ ] ] in conclusion , if @xmath216 is invariant by @xmath189 and @xmath219 , then@xmath240 \left ( \left [ x\right ] \right ) = \left\ { \begin{array } [ c]{l}\left [ g\right ] \left ( \left [ x+4\alpha\right ] \right ) \text { \ if } x\in(0,\pi-4\alpha)\\ \left [ g\right ] \left ( \left [ x+6\alpha\right ] \right ) \text { \ if } x\in\lbrack\pi-4\alpha,\pi ) .
\end{array } \right . \label{eq 31}\ ] ]
let @xmath265 such that @xmath266 .
observe that @xmath267 because @xmath264 . applying the invariance expressed in ( [ eq 31 ] )
iteratively , we have@xmath240 \left ( \left [ x\right ] \right ) = \left [ g\right ] \left ( \left [ x+\left ( 4k+6\right ) \alpha\right ] \right ) \text { for any } x\in\left ( 0,\pi\right )
. \label{eq 39}\ ] ] in other words , @xmath268 $ ] is invariant by the rotation of angle @xmath269 .
if @xmath270 is irrational , @xmath216 must be constant .
[ propositon characteristic function invariant]let @xmath271\right ) $ ] and suppose that @xmath272 is @xmath189-invariant .
if @xmath18 in ( [ eq probabilities model ] ) and ( [ eq tau maps ] ) is irrational , then @xmath273 equals either @xmath52 or @xmath226 .
suppose that @xmath193 .
we want to prove that @xmath272 can not be @xmath189-invariant .
since constants are trivially @xmath189-invariant , we can subtract @xmath273 from @xmath186 so that @xmath274 is still @xmath189-invariant and has expectation @xmath226 .
let @xmath275 and define @xmath276 , which is clearly @xmath219-invariant because @xmath277 . since @xmath195 is @xmath151-invariant , so are @xmath278 ( proposition [ prop symmetry ] ) and @xmath279 . by proposition [ prop g invariant s2 and phi ] , @xmath280 is constant and equal to its expectation @xmath281 $ ] .
but @xmath282 = \operatorname*{e}\nolimits_{\mu}\left [ g\right ] = 0 $ ] because @xmath283 and @xmath284 are invariant by @xmath219 .
therefore , @xmath285 = 0 $ ] and @xmath280 , which is constant , must be equal to @xmath226 .
that is,@xmath286 since@xmath287{l}1-\mu\left ( c\right ) > 0\text { if } x\in c\\ -\mu\left ( c\right ) < 0\text { if } x\notin c , \end{array } \right.\ ] ] ( remember that we assumed @xmath193 ) , we conclude from ( [ eq 25 ] ) that @xmath288 and@xmath289{l}g\left ( x\right ) = 1/2\longleftrightarrow x\in c,\\ g(x)=-1/2\longleftrightarrow x\in c^{c}. \end{array}\ ] ] moreover , looking carefully at the proof of proposition [ prop g invariant s2 and phi ] , we have@xmath240 \left ( \left [ x\right ] \right ) = \left\ { \begin{array } [ c]{l}\left [ g\right ] \left ( \left [ x+4\alpha\right ] \right ) \text { \ if } x\in(0,\pi-4\alpha)\\ -\left [ g\right ] \left ( \left [ x+6\alpha\right ] \right ) \text { \ if } x\in\lbrack\pi-4\alpha,\pi ) , \end{array } \right.\ ] ] where instead of @xmath245 we have now used @xmath290 . therefore , ( see ( [ eq 39]))@xmath240 \left ( \left [ x\right ] \right ) = -\left [ g\right ] \left ( \left [ x+\left ( 4k+6\right ) \alpha\right ] \right ) \text { for any } x\in(0,\pi ) , \label{eq 40}\ ] ] where @xmath267 is such that @xmath291 . the minus sign in ( [ eq 40 ] ) tells us that the rotation @xmath292 of angle @xmath293 sends @xmath294 to @xmath295 and vice versa .
let @xmath296 and @xmath297 the fourier expansions of @xmath186 and @xmath298 respectively . since @xmath299}$ ] , we have that@xmath300 on the other hand , we already argued that @xmath301 . imposing that the fourier coefficients of @xmath186 and @xmath298 are unique , we deduce that@xmath302 as @xmath303 is irrational , @xmath304 for any @xmath305 , which implies that @xmath306 for any @xmath305 so @xmath186 is constant a.s .. but this is clearly a contradiction because ( [ eq 25 ] ) implied @xmath307 .
in this section , we are going to prove that @xmath151 is exact .
this will imply knudsen s strong law for the random billiard introduced in section [ section model ] .
let @xmath308 be a measurable transformation of a probability space @xmath309 . from the measurability of @xmath79
, we have the chain of @xmath97-algebras@xmath310 the map @xmath79 is called * _ exact _ * if the @xmath97-algebra @xmath311 only contains sets of measure either @xmath226 or @xmath52 .
it is not difficult to prove that sets in @xmath312 are characterised by the property @xmath313 this characterisation , in turn , implies that a map @xmath79 is exact if and only if@xmath314 for any @xmath315 such that @xmath316 ( ( * ? ? ?
* section 2.2 ) ) .
equation ( [ eq 32 ] ) reads at the level of @xmath317 spaces as@xmath318 where @xmath319 and @xmath320 is the * _ koopman operator _ * defined on @xmath317 .
if a map is exact , then@xmath321 which means that @xmath322 $ ] in @xmath317 as @xmath323 ( see ( * ? ? ?
* section 2.5 ) ) .
if @xmath9 is a probability absolutely continuous with respect to @xmath61 with radon - nikodym derivative @xmath324 , then@xmath325 \underset{n\rightarrow \infty}{\longrightarrow}\mu\left ( a\right ) \operatorname*{e}\nolimits_{\mu } \left [ f\right ] = \mu(a),\ ] ] that is , the sequence of measures @xmath326 converges ( weakly ) to @xmath61 for any absolutely continuous measure @xmath179 ( ( * ? ? ?
* proposition 4.4.1 ( b ) ) , ( * ? ? ?
* section 2.6 ) ) .
we want to show that the skew - type representation @xmath327\rightarrow\lbrack0,1)\times\lbrack0,\pi]$ ] of the model introduced in section [ section model ] is exact with respect to the liouville measure to conclude that , for any absolutely continuous measure @xmath179 on @xmath328\right ) $ ] and any @xmath329)$ ] , @xmath330}\mathbf{1}_{a}\left ( \pi_{2}\circ s^{n}\right ) d\nu=\int_{[0,\pi]}u_{s}^{n}\left ( \mathbf{1}_{\pi_{2}^{-1}(a)}\right ) d\lambda\otimes\nu\underset{n\rightarrow\infty}{\longrightarrow } \mu\left ( a\right ) , \label{eq 33}\ ] ] i.e. , that the strong knudsen s law holds .
observe from ( [ eq 33 ] ) that we only need to consider the action of the koopman operator @xmath331 on square integrable functions that are the pull - back by @xmath332 of functions in @xmath333,\mathcal{b}\left ( [ 0,\pi]\right ) , \mu\right ) $ ] .
we are going to denote @xmath334,\mathcal{b}\left ( [ 0,\pi]\right ) , \mu\right ) $ ] simply by @xmath335\right ) $ ] for the sake of a clearer notation .
consequently , we do not need to check that @xmath151 is exact for the borel @xmath336-algebra of @xmath337 $ ] but it is enough to consider a smaller one , namely , the smallest @xmath97-algebra that makes both the functions in @xmath338\right ) \right ) $ ] and @xmath151 measurable .
for any @xmath339 , let @xmath340\right ) \right\ } \right ) , \ ] ] where @xmath168 are as in definition [ definition slices ] . for any @xmath341\right ) $ ] , the sets @xmath342 and @xmath343 will be called * generators * ( of @xmath344 ) .
the sequence of @xmath97-algebras @xmath345 define a filtration,@xmath346 indeed , for any @xmath341\right ) $ ] and any @xmath347 , a generator @xmath348 can be written as@xmath349 which implies @xmath350 .
we define @xmath351 as the limit of this filtration .
@xmath352 @xmath353\rightarrow\lbrack0,1)\times\lbrack0,\pi]$ ] is @xmath351-measurable . for any @xmath354\right )
$ ] , @xmath355 is also @xmath351-measurable .
the second statement of the proposition is obvious . to prove that @xmath151 is @xmath351-measurable it is enough to show that @xmath356 for the generators of @xmath351 .
so let @xmath357 such that @xmath358 where @xmath341\right ) $ ] and @xmath347 .
then , by definition of the sets @xmath168@xmath359 which clearly belongs to @xmath351 . the following two lemmas aim at getting a better insight of the structure of the @xmath97-algebra @xmath351 . as an immediate consequence of them
, it will be enough to show that @xmath151 is exact by looking at its action on the generators of @xmath351 .
[ prop finite union generators]any set @xmath360 can be expressed as a finite union of disjoint generators .
first to all , observe that , for a fixed @xmath339 , the sets @xmath168 are disjoint and form a partition of @xmath361 $ ] .
the complement of a generator @xmath362 is@xmath363 \backslash u\right ) , \ ] ] where @xmath364 denotes the disjoint union , so it can be expressed as a finite union of generators . on the other hand ,
if we take a countable union of generators @xmath365 , @xmath366,@xmath367 where @xmath368 is in @xmath369 .
but@xmath370 where the indices @xmath371 can only take finite number of possibilities .
therefore , a countable union of generators reduces to a finite union .
the union @xmath372 of a filtration @xmath373 is an algebra of sets .
the @xmath97-algebra generated by an algebra of sets @xmath374 is characterised by containing all the sets in @xmath374 and the limit of all monotone sequence of sets .
that is , @xmath375 coincides with the monotone class generated by @xmath374 ( ( * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* section 1.3 , theorem 1 ) ) .
this observation is the key to proving the following lemma : [ lemma a contains generator]if @xmath376 has positive probability , then @xmath94 contains a generator of positive probability .
let @xmath377 . if @xmath378 , then @xmath360 for some @xmath339 . by proposition [ prop finite union generators ]
, @xmath94 can be expressed as a finite union of generators and at least one of them must have positive probability . if @xmath379 , then there exists a monotone sequence of sets @xmath380 such that @xmath381 .
we are going to deal with the cases @xmath382 increasing or decreasing separately . *
if @xmath382 is increasing , then @xmath383 . since @xmath384 , at least one @xmath385 must have positive measure , @xmath386 . but
@xmath387 for some @xmath388 , so it contains a generator of positive probability .
* let @xmath389 be a decreasing sequence such that @xmath390 .
suppose that @xmath391 ( otherwise there is nothing to prove ) . we have@xmath392 so that @xmath94 can be expressed as the complement of a countable union of disjoint sets .
rename @xmath393 and @xmath394 , and express any @xmath395 as a disjoint union of generators ( lemma [ prop finite union generators]),@xmath396 in this decomposition , we only consider sets of strictly positive probability .
then,@xmath397 where @xmath398 .
we claim that there exists a generator @xmath399 of strictly positive probability that does not intersect @xmath400 , i.e. , @xmath401 .
+ let @xmath402 be a sequence appearing in the decomposition ( [ eq 38 ] ) . if @xmath403 , then@xmath404 \right\backslash \pi_{2}^{-1}\left ( \bigcup\nolimits_{m\in i_{i_{1}^{n} ...
i_{r}^{n}}}u_{m}\right ) \right ) \bigcap j_{i_{1}^{n} ... i_{r}^{n}}\ ] ] is a generator of strictly positive measure that does not intersect @xmath400 . if @xmath405 for any finite sequence @xmath402 in ( [ eq 38 ] ) , then @xmath406 a.s . and since we assumed that @xmath407 , there must exist some set @xmath408 with positive probability that does not appear in the decomposition of @xmath400 .
that is , @xmath409\right ) \bigcap j_{i_{1} ... i_{k}}\subseteq a$ ] . for any @xmath357 , @xmath410 in other words , @xmath353\rightarrow\lbrack0,1)\times \lbrack0,\pi]$ ] is an exact endomorphism of @xmath411,\mathcal{f},\lambda\otimes\mu\right ) $ ] .
let @xmath357 be such that @xmath412 . by lemma
[ lemma a contains generator ] , @xmath94 contains a generator @xmath413 of positive measure for some @xmath341\right ) $ ] , @xmath414 .
it is an immediate consequence of the definitions that , after @xmath13 iterations , the set @xmath415 has all its fibers of length @xmath52 , that is,@xmath416 let @xmath417 and let @xmath159 be the fiber at @xmath209 $ ] .
looking carefully at ( [ eq tau maps ] ) and ( [ eq probabilities model ] ) , one can see that@xmath418 where @xmath419 denotes the identity on @xmath25 $ ] , and@xmath420 these remarks imply that , if @xmath421 , @xmath422 since all the fibers of @xmath423 have length @xmath52 , @xmath424 and , iteratively,@xmath425 define @xmath426 .
obviously @xmath427 , the fibers of @xmath294 have all length @xmath52 , and @xmath428 .
we want to show that @xmath429 .
observe that @xmath430 because @xmath431 .
from @xmath427 we have@xmath432 but @xmath294 and @xmath433 have the same measure ( @xmath189 is @xmath59-preserving ) , so @xmath186 is a @xmath189-invariant function a.s .. since @xmath434 a.s .
( the fibers of @xmath294 have all length @xmath52 a.s . )
, @xmath435 has full measure by proposition [ propositon characteristic function invariant ] .
therefore , @xmath436 and @xmath437 this proves that @xmath189 is exact . in general , we have@xmath438 so that , using again that @xmath151 is @xmath59-preserving,@xmath439 which proves that @xmath440 is increasing and converges to @xmath52 , i.e. , @xmath151 is also exact . to sum up , the skew - type representation @xmath441\rightarrow\lbrack0,1)\times\lbrack0,\pi]$ ] associated to the random map ( [ eq tau maps ] ) and ( [ eq probabilities model ] ) is exact which implies , by ( [ eq 33 ] ) , that @xmath442 as @xmath323 for any initial distribution @xmath9 on @xmath25 $ ] absolutely continuous with respect to the law @xmath443 , @xmath444\right ) $ ] . since @xmath445 on @xmath446 , @xmath61 and
the lebesgue measure @xmath60 are absolutely continuous with respect to each other on @xmath447 . in other words ,
_ strong knudsen s law _
@xmath442 _ holds for any initial distribution absolutely continuous with respect to _ @xmath448__. _ _
in this section , we are going to show that numerical simulations are in accordance with theoretical results . with this aim
, we take an arbitrary initial distribution on @xmath230 $ ] such as in the following picture and we make it evolve according to our random system ( [ eq tau maps ] ) and ( [ eq probabilities model ] ) . initial.jpg + initial normalised distribution of particles .
[ fig initial distribution ] since we can only simulate a finite number particles in a computer , in this experiment we take a total amount of 30000 balls and divide @xmath449 $ ] in subintervals of the same length ( 45 in our experiment ) .
that is , we approximate the initial distribution by a step function . in each subinterval
, we put the proportion of balls according to the probability density function above .
the initial angle associated to any of those balls is the middle point of the interval where they fall .
the simulation then goes as follows . at the @xmath121-_th _ step ,
we take the @xmath450-_th _ particle with angle @xmath451 and a random number @xmath452 uniformly distributed between @xmath226 and @xmath52 , one for each particle . if @xmath453 , then @xmath454 , where @xmath455 are as in ( [ eq tau maps ] ) , and so on . after a long number of iterations , the distribution stabilises as follows : final.jpg + final normalised distribution of particles . in red , @xmath456 .
[ fig final distribution ] as we can see , the outline of the final distribution tends to the graph of @xmath457 .
the small inaccuracy is explained by the fact that only a finite ( i.e. , a discrete ) number of initial angles is considered .
the smaller the subintervals in which @xmath230 $ ] is divided are , the better the final distribution approximates @xmath457 .
we have repeated this experiment over several initial distributions obtaining always similar results , which experimentally confirms the validity of knudsen s law for our model .
[ proof of lemma [ lemma slices go to zero]]to start with , will prove that @xmath458 from the very definition of @xmath151 and @xmath155 , we have@xmath459 on the other hand , let @xmath460 .
then @xmath461 for some @xmath462 .
let @xmath463 .
it is obvious that @xmath464 as @xmath465 because @xmath466 is not empty .
moreover , @xmath467 .
therefore @xmath468 and@xmath469 identifying both fibers @xmath466 and @xmath470 as subsets of @xmath471 , the map@xmath289{rcl}s_{x}:i_{x}\bigcap j_{i_{1}\cdots i_{n}}\subset\lbrack0,1 ) & \longrightarrow & i_{\tau_{i_{n}}(x)}\bigcap j_{i_{1}\cdots i_{n-1}}\\ y & \longmapsto & \varphi_{i_{n}}(y,\tau_{i_{n}}(x))=\frac{1}{p_{i_{n}}(x)}\left ( y-{\textstyle\sum\nolimits_{j=1}^{i_{n}-1 } } p_{j}(x)\right ) , \end{array } \label{eq 37}\ ] ] is then a diffeomorphism . after having identified @xmath472 as a subset of @xmath471 , we can compute its lebesgue measure @xmath473 .
consequently,@xmath474 applying iteratively ( [ eq 19 ] ) , we obtain@xmath475 n. chernov and r. markarian . _
introduction to the ergodic theory of chaotic billiards .
_ monografas del instituto de matemtica y ciencias afines * 19*. instituto de matemtica y ciencias afines ; pontificia universidad catlica del per , lima , 2001 . | we prove knudsen s law for a gas of particles bouncing freely in a two dimensional pipeline with serrated walls consisting of irrational triangles .
dynamics are randomly perturbed and the corresponding random map studied under a skew - type deterministic representation which is shown to be ergodic and exact . * keywords : * knudsen s law , random billiards , random maps , irrational polygons , ergodic billiards . |
this work was supported by basic science research program through the national research foundation of korea ( nrf ) funded by the ministry of education , science and technology ( 2010- 0013054 ) ( swk ) , ( 2 - 2013 - 4569 - 001 - 1 ) ( mip ) .
v. e. hubeny , phys .
d * 59 * , 064013 ( 1999 ) [ gr - qc/9808043 ] ; t. jacobson and t. p.
sotiriou , phys .
lett . * 103 * , 141101 ( 2009 ) [ phys .
lett . * 103 * , 209903 ( 2009 ) ] [ arxiv:0907.4146 [ gr - qc ] ] ; e. barausse , v. cardoso and g. khanna , phys .
lett . * 105 * , 261102 ( 2010 ) [ arxiv:1008.5159 [ gr - qc ] ] ; m. colleoni and l. barack , phys .
d * 91 * , 104024 ( 2015 ) [ arxiv:1501.07330 [ gr - qc ] ] .
g. e. a. matsas and a. r. r. da silva , phys .
lett . * 99 * , 181301 ( 2007 ) [ arxiv:0706.3198 [ gr - qc ] ] ; s. hod , phys .
* 100 * , 121101 ( 2008 ) [ arxiv:0805.3873 [ gr - qc ] ] ; m. richartz and a. saa , phys . rev .
d * 84 * , 104021 ( 2011 ) [ arxiv:1109.3364 [ gr - qc ] ] ; s. aretakis , commun .
phys . * 307 * , 17 ( 2011 ) [ arxiv:1110.2007 [ gr - qc ] ] ; annales henri poincare * 12 * , 1491 ( 2011 ) [ arxiv:1110.2009 [ gr - qc ] ] .
e. poisson and w. israel , phys .
d * 41 * , 1796 ( 1990 ) .
d. marolf , gen .
. grav . * 42 * , 2337 ( 2010 ) [ arxiv:1005.2999 [ gr - qc ] ] .
h. a. shinkai and s. a. hayward , phys .
d * 66 * , 044005 ( 2002 ) [ gr - qc/0205041 ] .
m. i. park , phys .
b * 647 * , 472 ( 2007 ) [ hep - th/0602114 ] ; phys .
b * 663 * , 259 ( 2008 ) [ hep - th/0610140 ] ; class .
grav . * 25 * , 095013 ( 2008 ) [ hep - th/0611048 ] ; phys .
d * 77 * , 026011 ( 2008 ) [ hep - th/0608165 ] ; phys .
d * 77 * , 126012 ( 2008 ) [ hep - th/0609027 ] . | there have been lots of debates about the final fate of an evaporating black hole and the singularity hidden by an event horizon in quantum gravity .
however , on general grounds , one may argue that a black hole stops radiation at the planck mass @xmath0 , where the radiated energy is comparable to the black hole s mass . and also , it has been argued that there would be a _ wormhole - like _ structure , known as space - time foam " , due to large fluctuations below the planck length @xmath1 . in this paper , as an explicit example , we consider an exact classical solution which represents nicely those two properties in a recently proposed quantum gravity model based on different scaling dimensions between space and time coordinates .
the solution , called black wormhole " , consists of two different states , depending on its mass parameter @xmath2 and an ir parameter @xmath3 : for the black hole state ( with @xmath4 ) , a _
non - traversable _ wormhole occupies the interior region of the black hole around the singularity at the origin , whereas for the wormhole state ( with @xmath5 ) , the interior wormhole is exposed to an outside observer as the black hole horizon is disappeared from evaporation .
the black hole state becomes thermodynamically stable as it approaches to the merge point where the interior wormhole throat and the black hole horizon merges , and the hawking temperature vanishes at the exact merge point ( with @xmath6 ) .
this solution suggests the generalized cosmic censorship " by the existence of a wormhole - like structure which protects the naked singularity even after the black hole evaporation .
one could understand the would - be wormhole inside the black hole horizon as the results of microscopic wormholes created by negative " energy quanta which have entered the black hole horizon in hawking radiation processes ; _ the quantum black hole could be a wormhole factory ! _
it is found that this speculative picture may be consistent with the recent @xmath7 " proposal for resolving the black hole entanglement debates .
it is widely accepted that general relativity ( gr ) would not be appropriate for describing the small scale structure of space - time .
for example , gr , when combined with quantum mechanics , provides a length scale @xmath8 , which may provide an absolute limitation for the measurements of space - time distances sale .
actually , this is the length scale on which quantum fluctuations of the space - time are expected to be of order of unity . on the other hand , the singularity theorem , stating the necessary existence of singularities , where the classical concept of space and time breaks down , at certain space - time domains with some reasonable assumptions in gr penr ,
may be regarded as an indication of the incompleteness of gr .
these circumstances may provide strong motivations to find the quantum theory of gravity which can treat the above mentioned problems of gr .
actually , the necessity of quantizing the gravity has been argued in order to have a consistent interaction with a quantum system eppl .
moreover , it has been also shown that even small quantum gravitational effects dramatically change the characteristic features of a black hole so that it can emit radiations though the causal structures of the classical geometry is unchanged in the semiclassical treatment hawk . however , as the black hole becomes smaller and smaller by losing its mass from emitting particles , the semiclassical treatment becomes inaccurate and one can not ignore the back reactions of the emitted particles on the metric and the quantum fluctuations on the metric itself anymore .
actually , regarding the back reaction effects , one can argue that a black hole stops radiations at the planck mass @xmath9 , where the radiated energy is comparable to the black hole s mass , since a black hole can not radiate more energy than it has , via the pair creation process near the black hole horizon .
this implies that the black hole should become thermodynamically stable as it becomes smaller and finally has the vanishing hawking temperature at the smallest black hole mass .
it seems that this should be one of verifiable predictions that any theory of quantum gravity make bona .
moreover , according to large fluctuations of metric below the planck length @xmath10 sale , the _ wormhole - like _ structure , known as space - time foam " , has been proposed by wheeler whee .
this may be another verifiable prediction of the quantum gravity also .
the purpose of this paper is to consider an exact classical solution , as an explicit example , which represents nicely those two properties in a recently proposed quantum gravity model , known as hoava gravity , based on different scaling dimensions between space and time coordinates .
the solution , called black wormhole " , consists of two different states depending on its mass parameter @xmath2 and an ir ( infrared ) parameter @xmath3 : for the black hole state ( with @xmath4 ) , a _
non - traversable _ wormhole occupies the interior region of the black hole around the singularity at the origin , whereas for the wormhole state ( with @xmath5 ) , the interior wormhole is exposed to an outside observer as the black hole horizon is disappeared from evaporation .
the black hole state becomes thermodynamically stable as it approaches to the merge point where the interior wormhole throat and the black hole horizon merges , and the hawking temperature vanishes at the exact merge point ( with @xmath6 ) .
the solution suggests that , in quantum gravity , the ` conventional ' cosmic censorship can be generalized even after black hole evaporation by forming a wormhole throat around the used - to - be singularity . in gr , black hole and wormhole
are quite distinct objects due to their completely different causal structures .
but the claimed generalized cosmic censorship " suggests that the end state of a black hole is a wormhole , not a naked singularity .
this may correspond to a foam - like nature of space - time at short length scales .
furthermore , one could understand the would - be wormhole inside the black hole horizon as the results of microscopic wormholes created by negative " energy quanta which have entered the black hole horizon in hawking radiation processes so that _ the quantum black hole could be a wormhole factory .
_ it is found that this speculative picture may be consistent with the recent @xmath7 " proposal for resolving the black hole entanglement debates . to see how this picture can arise explicitly , we consider the hoava gravity which has been proposed as a four - dimensional , renormalizable , higher - derivative quantum gravity without ghost problems , by adopting different scaling dimensions for space and time coordinates in uv ( ultraviolet ) energy regime , @xmath11=-1 , [ { \bf x}]=-z$ ] with the dynamical critical exponents @xmath12 , at the expense of lorentz invariance " hora .
we formally define the quantum gravity by a path integral z = e^is/ with the proposed action ( @xmath13 is considered , for simplicity ) , up to the surface terms , s & = & dt d^3 x n , [ horava ] and with the adm decomposition of the metric , ds^2=-n^2 c^2 dt^2+g_ij(dx^i+n^i dt)(dx^j+n^j dt ) , the extrinsic curvature , k_ij=(_ij-_i n_j-_jn_i ) , the cotton tensor , c^ij=^ik_k ( r^(3)j_-r^(3 ) ^j _ ) , and coupling constants , @xmath14 .
the last term in the action ( [ horava ] ) represents a soft " violation , with the ir parameter @xmath3 , of the detailed balance " condition in hora and this modifies the ir behaviors so that newtonian gravity limit exists keha , park , kiri .
the proposed action is not the most general form for a power - counting renormalizable gravity , compatible with the assumed foliation preserving _ diff _ but it is general enough to contain all the known gr solutions , and the qualitative features of the solutions are expected to be similar lu , keha , park , kiri . here , originally ,
the _ non - relativistic _ higher - derivative deformations were introduced from the technical reason of the necessity of renormalizable interactions without the ghost problem which exists in relativistic higher - derivative theories hora .
but we further remark that , which has not been well emphasized before , the ( uv ) lorentz violation _
might _ have a more fundamental reason in our quantum gravity set - up since this may be consistent with the existence of the absolute minimum length @xmath10 which does not depend on the reference frames , violating the usual _ relativistic _ length contraction . in hoava gravity for varying @xmath2 with a fixed @xmath3 ( left ) and for varying @xmath3 with a fixed @xmath2 ( right ) . in particular ,
in the left we consider @xmath15 ( top to bottom solid curves ) with @xmath16 , and in the right @xmath17 ( top to bottom solid curves ) with @xmath18 , in contrast with schwarzschild solution for @xmath18 in gr ( dotted curve).,title="fig:",width=275 ] in hoava gravity for varying @xmath2 with a fixed @xmath3 ( left ) and for varying @xmath3 with a fixed @xmath2 ( right ) .
in particular , in the left we consider @xmath15 ( top to bottom solid curves ) with @xmath16 , and in the right @xmath17 ( top to bottom solid curves ) with @xmath18 , in contrast with schwarzschild solution for @xmath18 in gr ( dotted curve).,title="fig:",width=275 ] for the simplest case of static , _
i.e. _ , non - rotating , uncharged black holes , where only the last three terms in the action ( [ horava ] ) are relevant , the exact solutions have been found completely for arbitrary values of coupling constants , @xmath19 , and @xmath3 lu , keha , park , kiri .
however , for the present purpose we only consider a simple example of @xmath20 , ds^2=-n(r)^2 c^2 dt^2++r^2 ( d^2+^2d^2 ) [ bh1 ] with n^2=f=1+r^2- [ bh2 ] so that the standard einstein - hilbert action and the asymptotically flat , schwarzschild black hole solution are recovered in the ir limit , _
i.e. _ , @xmath21 with @xmath22 keha , park . here
@xmath2 denotes the ` @xmath23 ' and the positive ir parameter @xmath3 controls the strength of higher - derivative corrections so that the limit @xmath24 with ` @xmath25 ' corresponds to gr limit .
one remarkable property of the solution is that there is an inner horizon @xmath26 as well as the outer horizon @xmath27 , which solves @xmath28 , at r _ = m ( 1 ) as the result of higher ( spatial ) derivatives ( fig .
1 ) : the higher derivative terms act like some ( non - relativistic ) effective matters in the conventional einstein equation so that there is some repulsive " interaction at short distances . moreover , even though the metric converges to the minkowski s flat space - time at the origin @xmath29 , _
i.e. _ , @xmath30 , its derivative is not continuous so that there is a curvature singularity at @xmath29 , which may be captured by the singularity of @xmath31 .
so , even though the singularity at @xmath29 is a time - like line ( _ i.e. _ , time - like singularity ) and is milder than that of schwarzschild black hole , @xmath32 ( and also reissner - nordstrm s @xmath33 ) and surrounded by the ( two ) horizons provided m^2 12 , this might indicate that the proposed gravity does not completely resolve the singularity problem of gr still , classically .
this circumstance looks inconsistent with the cosmology solution , where the initial singularity does not exist whence there exist the higher - derivative effects , _ i.e. _ , non - flat universes keha , park .
moreover , the black hole singularity becomes naked for @xmath5 so that the cosmic censorship might not work in this edge of solution space , even if @xmath2 is positive definite .
but according to recent bottacantcheff - grandi - sturla ( bgs ) s construction of wormholes , it seems that there is another possibility for resolving the unsatisfactory circumstance ca nt .
what they found was that there exits a wormhole solution also , in addition to the naked singularity solution for @xmath5 , without introducing additional ( exotic ) matters at the throat .
their obtained wormhole solution ds^2=-n_(r)^2 c^2 dt^2++r^2 ( d^2+^2d^2 ) [ wormhole ] with n_^2=f_=1 + _ r^2- [ wormhole2 ] is made of two coordinate patches , each one covering the range @xmath34 in one universe and the two patches joining at the wormhole throat @xmath35 , which is defined as the minimum of the radial coordinate @xmath36 . in the conventional approach for ( traversable ) wormholes , there are basically two unsatisfactory features morr , viss .
first , we do not know much about the mechanism for a wormhole formation .
this is in contrast to the black hole case , where the gravitational collapse of ordinary matters , like stars , can form a black hole , if its mass is enough .
second , the usual steps of ` constructing ' wormholes are too artificial as summarized by the following three steps : ( 1 ) .
prepare two or several universes ; ( 2 ) .
connect the universes by cut and paste of their throats ; ( 3 ) .
put the needed
exotic " matters , which violate the energy conditions , to the throats so that einstein s equations are satisfied .
however , in the new approach , the problem of artificiality of a wormhole construction is avoided by observing that the solution is smoothly joined at the throat , so that the additional compensating ( singular or non - singular ) matters are not needed , if the metric and its derivatives are continuous at the throat .
but , if we further consider the reflection symmetric two universes , _
i.e. _ , @xmath37 , with @xmath38 , the only possible way of smooth patching at the throat is ) is not always necessary for @xmath39 but more general class of wormhole solutions could be possible .
but , this is only for the case of singular , @xmath40-function discontinuities in the equations of motions .
whereas , for the non - singular discontinuities at the throat , which can not be properly treated in the bgs s analysis , the condition ( [ junction ] ) is still essential for our wormhole construction . ]
.df_dr|_r_0 = 0 [ junction ] and the throat radius is obtained as r_0=(m2 ) ^1/3 . here , it is important to note that the throat is located always inside the black hole horizon , _
i.e. _ , @xmath41 for @xmath4 so that it is unobservable to an outer observer , whereas the wormhole throat emerges for @xmath5 , instead of a naked singularity ( fig .
2 ) . for a fixed @xmath3 , the throat radius , after emerging from the coincidence with the extremal black hole radius , @xmath42 , decreases monotonically as @xmath2 decreases and finally vanishes for @xmath43 , _
i.e. _ , minkowski vacuum ( fig .
2 , left ) .
this is the situation that has been assumed in bgs s paper ca nt .
but , on the other hand , for a fixed @xmath2 and varying @xmath3 , one finds that the throat radius increases again indefinitely as @xmath3 decreases .
this means that the wormhole size can be quite large when the coupling constant @xmath3 , which could flow under renormalization group , becomes smaller at quantum gravity regime , like the planck size black hole or wormhole ( fig .
2 , right ) .
cai , rather than another form of an integration constant @xmath44 lu , park ] in the actual quantum gravity process , like black hole evaporations , @xmath2 , due to hawking radiations , as well as @xmath3 , due to renormalization group flow with the change of energy scales , can vary so that we need to consider some combinations of situations of the left and right in fig .
2 , by extending the original interpretation . ) and bottom ( @xmath45 ) curves ) and wormhole throat ( middle ( @xmath35 ) curve ) radii for varying @xmath2 with a fixed @xmath16 ( left ) , and for varying @xmath3 with a fixed @xmath18 ( right).,title="fig:",width=275 ] ) and bottom ( @xmath45 ) curves ) and wormhole throat ( middle ( @xmath35 ) curve ) radii for varying @xmath2 with a fixed @xmath16 ( left ) , and for varying @xmath3 with a fixed @xmath18 ( right).,title="fig:",width=275 ] so , the wormhole solution is obtained , without the problem of artificiality , when the black hole horizon disappears for @xmath5 . whereas , the solution for @xmath4 with the throat inside the horizon is _ not _ the traversable " wormhole since the throat will be still inside the black hole horizon of the mirror " black hole in another universe .
this is similar to the einstein - rosen bridge eins but the difference is that the throat may not coincide with the black hole horizon generally in the present case .
moreover , in the present case the throat is located at the fixed time " @xmath35 so that any time - like trajectories should meet the throat if exits ( fig .
3 ) .
this means that the black hole solution ( [ wormhole ] ) , ( [ wormhole2 ] ) with the time - like " throat for @xmath4 should be considered as a physically distinct object from the black hole solution ( [ bh1 ] ) , ( [ bh2 ] ) having the ( hidden ) singularity at @xmath29 . and
in order to avoid a rather strange situation that the wormhole throat suddenly " emerges from the extremal black hole which has a singularity at @xmath29 still , it would be _ natural _ to consider the time - like throat in ( [ wormhole ] ) , ( [ wormhole2 ] ) for @xmath4 as the would - be " wormhole throat . in order
to be distinguished from the usual black hole solution ( [ bh1 ] ) , ( [ bh2 ] ) , we may call the solution ( [ wormhole ] ) , ( [ wormhole2 ] ) as the black wormhole " solution .
inside the black hole horizons @xmath27 .
the arrows represent the time - like geodesics inside the horizons .
, width=453 ] now , we have a completely regular vacuum solution , without the curvature singularities , which interpolates between the black hole state for @xmath4 and wormhole state for @xmath5 .
this seems to support wheeler s foam picture of the quantum space - time in quantum gravity but as a real static , not as a virtual time - dependent , wormhole .
and in our quantum gravity theory , where the concept of horizon emerges at low energies , escaping the horizons is not impossible for high energy particles with lorentz - violating dispersion relations .
but for low energy point of view , without probing the interior structure of real black holes , the observable consequences of the black wormhole solution would be expressed in the form of a
generalized cosmic censorship " , suggesting that the naked singularity does not appear still by forming wormholes even after the horizons disappear in quantum gravity " though , before evaporation , there would be no naked singularity by the existence of horizons as usual .
another important new implication of the black wormhole solution to low energy observers is that there would be transformations between black holes and wormholes , which is known to be impossible in gr , due to the no - go theorem for topology change gero . in other words , once a ( primordial ) wormhole is formed in the quantum gravity regime , due to quantum fluctuation in the early universe , it may evolve into a black hole state by the combinations of fig .
2 ( right ) from renormalization group flows to gr and fig . 2 ( left ) from accretion of matters
. it would be interesting to see whether this could be a mechanism for the primordial black holes and supermassive black holes , which are believed to be formed very early in the universe and distinguished from the stellar - mass black holes which are ( or believed to be ) generated from collapsing stars . on the other hand
, according to the black wormhole solution , once a black hole is formed in our quantum gravity , it always has the would - be wormhole inside the horizon but this inside wormhole is exposed to outer observers when the horizon disappears after the complete evaporation .
but we know that the inside wormhole is absent in the gr limit , as can be seen in fig .
2 ( right ) . then , where does the inside wormhole come from in quantum gravity regime ? this is the question about the physical mechanism of a wormhole formation , which has been lacking .
it seems that the only possible answer to this question could be found in the hawking radiation process , which involves virtual pairs of particles near the event horizon , one of the pair enters into the black hole while the other escapes : the escaped particle is observed as a real particle with a _ positive _ energy with respect to an observer at infinity and then , the absorbed particle must have a _ negative _ energy in order that the energy is conserved hawk .
this implies that the negative energy particles that fall into black hole could be an
exotic matter " source for a wormhole formation inside the horizon .
but one can easily discover that this would be quite implausible in gr since negative masses ( or energies ) repel ( _ i.e. _ , produce the outward accelerations of ) , as positive masses attract ( _ i.e. _ , produce the inward accelerations of ) , all other bodies regardless of their ( positive or negative ) masses from the _ equivalence principle_. we have never observed the negative mass object yet and it could produce some strange phenomena , known as the runaway motion " of a pair of positive and negative mass particles , but the repulsive nature of gravity for negative masses is another remarkable consequence of gr bond . in the language of the newtonian approximation , where the gravitational potential of a spherically symmetric body is given by @xmath46 , that property is indicated by the fact that @xmath47 , which is related to the ( radial ) acceleration @xmath48 , is always positive for the potential of a positive mass @xmath49 , whereas @xmath47 is always negative for the potential of a negative mass @xmath50 .
this property implies that negative mass could not form a ( stable ) structure , _
naturally_. this could explain why it would not be possible to form a wormhole by collapsing of exotic matters , in particular , negative energy matters , including the negative energy particles that fall into black hole in gr , in contrast to the case of a black hole formation . actually , previously the negative energy particle has been thought to result in just reducing the black hole mass by some compensation process with positive energy sources , which may exist inside the horizon , for the positive black hole mass .
however , the situation is quite different in our quantum gravity context , where the gravity becomes weaker at short distance and changes its ( attractive or repulsive ) nature after passing the surface of @xmath51 , which we call the zero - gravity surface " for convenience . in other words , for the positive black hole mass @xmath49
, the gravity becomes repulsive inside the zero - gravity surface ( fig .
1 , solid curves ) .
this now implies that even the positive mass could not form a structure _ inside the zero - gravity surface _ , naturally , in contrast to the outside the zero - gravity surface where the its gravity is still attractive so that a stable structure can be formed , as in gr . in hoava gravity for varying @xmath2 with a fixed @xmath3 ( left ) and for varying @xmath3 with a fixed @xmath2 ( right ) . in particular , in the left we consider @xmath52 ( bottom to top solid curves ) with @xmath53 , and in the right @xmath54 ( bottom to top solid curves ) with @xmath55 , in contrast with the negative mass schwarzschild solution for @xmath55 in gr ( dotted curve).,title="fig:",width=275 ] in hoava gravity for varying @xmath2 with a fixed @xmath3 ( left ) and for varying @xmath3 with a fixed @xmath2 ( right ) . in particular , in the left we consider @xmath52 ( bottom to top solid curves ) with @xmath53 , and in the right @xmath54 ( bottom to top solid curves ) with @xmath55 , in contrast with the negative mass schwarzschild solution for @xmath55 in gr ( dotted curve).,title="fig:",width=275 ] in order to see how this problem could be resolved dramatically by the negative masses in our quantum gravity context , we now turn to consider the _ negative mass _ solution in hoava gravity , which has never been studied in the literature . in order to find a negative mass solution
, one might first try to consider @xmath50 for the solution ( 7 ) but one can easily find that this could not be the right solution : the solution is _ not defined _ below a certain radius , where the quantity inside the square root become zero for @xmath56 or shows a different asymptotic ( @xmath57 ) behavior , _
i.e. , _ not asymptotically flat , @xmath58 , for @xmath59 . however , by noting that the solution of the metric ansatz ( 6 ) is unique up to the ( @xmath60 ) sign in front of the square root in ( 7 ) generally , we find that the negative mass solution as n^2=f=1+r^2 + " [ negative_m ] for the different sign of the square root term , compared to the solution ( 7 ) .
this solution recovers the schwarzschild solution in the ir limit for @xmath59 , as ( 7 ) does for @xmath49 and @xmath56 ( fig .
4 ) ; for @xmath56 , the solution ( 14 ) neither recovers the schwarzschild solution nor the metric is defined for the whole space region . [
but in this case , in order that gr is recovered , we need to consider @xmath61 as in asymptotically de sitter space park and we will discuss about this again later ] one can prove that the new solution ( 14 ) does not have horizons , as can be easily observed in fig . 4 also but it has a curvature singularity at @xmath29 , where the metric is finite but discontinuous , with the same degrees of divergences as the solution ( 7 ) .
so , this solution itself represents a naked singularity as the negative mass schwarzschild solution does in gr .
actually , this may be one of main obstacle for considering the negative mass solution as a viable solution in gr and this is not much improved for the solution ( 14 ) alone , in our hoava gravity context .
but , as in the positive mass case , we have another ( exact ) solution of a wormhole , for the reflection symmetric two universes , with the same formula ( 13 ) for the throat @xmath35 .
in other words , in hoava gravity , there exists a regular , _
i.e. , singularity free _ , negative mass solution so that it could be viable but only in the form of a wormhole geometry .
but here we will not consider about the negative mass wormhole solution further since its formation mechanism is unclear at present .
rather , we consider the negative mass solution ( 14 ) in order to see how does the negative masses , like the negative energy particles that fall into black hole could interact and form a structure , like the ( black ) wormhole with a positive mass . actually , the solution ( 14 ) indicates that the gravity of a negative mass is repulsive at large distance but becomes weaker , _
i.e. _ , less - repulsive , at short distance and moreover becomes attractive inside the zero - gravity surface of @xmath51 .
this implies that the negative masses could form a structure naturally , at short distance inside the zero - gravity surface , in contrast to the case of positive masses which could not form a structure inside their _ own _ zero - gravity surface as we have explained already .
this is the remarkable consequence of hoava gravity which could justify our picture that _ wormholes are created and sustained by the continuous inflow of negative energy particles via the hawking radiation process _ , which is impossible in gr .
this is the main claim of this paper and in a more compact form , this can be expressed as : _ the quantum black hole could be a wormhole factory ! _ interestingly , this may provide a physical origin of the einstein - rosen bridge in the recent @xmath7 " proposal for resolving the issue of entanglement in a black hole spacetime mald claiming that the black hole and its hawking radiation are entangled via einstein - rosen bridges " ( fig . 5 ) . following their proposal , one could understand the would - be wormholes inside the black hole horizon as the results of microscopic wormholes created by negative energy quanta which have entered the black hole horizon but entangled with its hawking radiated , positive energy , partner quanta , which now lives in a single universe , widely separated from its mother black hole
. this may be consistent with our reflection symmetric configuration , but it is not clear how to extend this picture to more general configurations which can not be interpreted as a configuration in a single universe .
moreover , it is not clear either how the throats of einstein - rosen s bridges evolve into our would - be wormhole in detail . in conclusion , we have considered a vacuum and static _ black wormhole _ solution which is regular , _
i.e. _ , singularity - free , and interpolates between the black hole state for @xmath4 and wormhole state for @xmath5 through the coincidence state of an extremal black hole and a ( kind of ) einstein - rosen bridge for @xmath6 . from this
, we have suggested the transformation between the black hole and wormhole states , and its resulting generalized cosmic censorship . and
furthermore , we have argued that the would - be wormholes inside the black hole horizon could be understood as the results of microscopic wormholes created by negative " energy quanta which have entered the black hole horizon in hawking radiation processes so that _ the quantum black hole could be a wormhole factory . _ but then , can the transformation really occur dynamically ? " in order to answer to this question , we first consider the transformation from a black hole state to a wormhole state .
of course , there is some obstacle in gr context since the black hole should pass the extremal black hole to become a wormhole state but the extremal black hole has vanishing hawking radiation and temperature so that it is believed as the stable ground state in the black hole states .
however recently , certain classical instability , known as _ aretakis instability _ , for extremal black holes have been found in gr so that the usual belief may not be quite correct aret . a heuristic argument
about this instability suggests that the inner horizon instability of near - extremal black holes pois might cause an instability of the coincided inner and outer horizons maro . if this is the case , we may expect a similar instability in our extremal black holes also , due to the ( expected ) inner horizon instability .
park:2006 , which seems to be a quite generic property of the inner horizons , including those of the black holes in hoava gravity park .
] now then , let us consider the transformation from a wormhole state to a black hole state . in this case , there does not seem to exist any classical obstacle against this transformation , which resembles the collapse of stars made of ordinary matter to make a black hole , where the entropy would increase in the process . actually , in the conventional wormhole context with exotic matters in gr , the collapse of the morris - thorne wormhole into a black hole was observed from non - linear instability under ordinary matter perturbations shin .
we expect the similar non - linear instability , though linearly stable , exits for our wormhole state too so that it transforms to the extremal black hole state , a seed of large black holes . finally , two further remarks are in order .
first , even though we have obtained the solution in a particular quantum gravity model which is power - counting renormalizable without ghost problems , the features of the small scale structure seems to be quite generic if _ the vanishing hawking temperature for a planck mass black hole _ , which implying the existence of the wormhole throat @xmath35 satisfying @xmath62 , is considered as a verifiable prediction that any theory of quantum gravity makes . for example , so called , renormalization group improved black hole spacetimes " would have the similar wormhole structure and our discussion may not be limited to hoava gravity bona .
but we can easily check that it is not applicable to kerr nor reissner - nordstrm black holes in gr : in these cases , there are more metric functions or additional matter fields but there is no solution for the throat where all the metric functions or fields join smoothly .
second , we have found that , in order to describe the negative mass solution ( 14 ) , we need to consider the coupling constant of @xmath61 so that one can recover the well - defined gr parameters @xmath63 in the ir limit .
this does not affect the couplings of the second - derivative terms in the action ( 2 ) ( note that we are considering the case of @xmath64 ) but only affects other uv couplings lu , which differ from those for the positive mass solution ( 7 ) .
this indicates that the ( weak ) _ equivalence principle _ may be violated by uv effects generally , depending on the sign of particle s mass .
it would be a challenging problem that the runaway motions of a pair of positive and negative mass particles could be avoided from the inequivalent motions of the positive and negative mass particles in uv . |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.